MiDAS tutorial

Maciej Migdał & Christian Hammer

2022-11-01

logo

Introduction

Welcome to MiDAS. This tutorial is supposed to help you get started with your analyses of immunogenetic associations. We will work with a simulated data set of 500 patients and 500 controls with a binary disease diagnosis.

We also have high resolution HLA alleles (4 - 6 digit), and presence/absence calls for KIR genes.

Data import and sanity check

First, let’s import the phenotype data and HLA calls using MiDAS import functions. MiDAS will check for correct nomenclature of HLA. We can also define 4-digit resolution for HLA alleles as import format, which means that alleles with higher resolution will be reduced.

dat_pheno <-
  read.table(
  file = system.file("extdata", "MiDAS_tut_pheno.txt", package = "midasHLA"),
  header = TRUE
  )
dat_HLA <-
  readHlaCalls(
  file = system.file("extdata", "MiDAS_tut_HLA.txt", package = "midasHLA"),
  resolution = 4
  )

Let’s take a look at the imported HLA data tables:

HLA data as imported by MiDAS
ID A_1 A_2 B_1 B_2 C_1 C_2 DPA1_1 DPA1_2 DPB1_1 DPB1_2 DQA1_1 DQA1_2 DQB1_1 DQB1_2 DRA_1 DRA_2 DRB1_1 DRB1_2
C001 A*25:01 A*26:01 B*07:02 B*18:01 C*12:03 C*07:02 DPA1*01:03 DPA1*01:03 DPB1*02:01 DPB1*04:01 DQA1*05:05 DQA1*01:02 DQB1*06:02 DQB1*03:01 DRA*01:02 DRA*01:02 DRB1*15:01 DRB1*12:01
C002 A*02:01 A*02:324 B*50:01 B*18:01 C*06:02 C*12:03 DPA1*01:03 DPA1*01:03 DPB1*04:02 DPB1*04:02 DQA1*02:01 DQA1*05:05 DQB1*02:02 DQB1*03:01 DRA*01:01 DRA*01:01 DRB1*07:01 DRB1*11:04
C003 A*24:02 A*24:04 B*46:01 B*40:06 C*01:03 C*15:02 DPA1*02:02 DPA1*02:01 DPB1*05:01 DPB1*14:01 DQA1*01:04 DQA1*01:03 DQB1*06:01 DQB1*05:02 DRA*01:02 DRA*01:01 DRB1*14:07 DRB1*08:03
C004 A*01:01 A*24:02 B*08:01 B*15:01 C*07:01 C*03:03 DPA1*01:03 DPA1*01:03 DPB1*04:01 DPB1*03:01 DQA1*01:03 DQA1*01:02 DQB1*06:04 DQB1*06:03 DRA*01:02 DRA*01:01 DRB1*13:01 DRB1*13:02
C005 A*01:01 A*25:01 B*18:01 B*08:01 C*12:03 C*07:01 DPA1*01:03 DPA1*01:03 DPB1*04:01 DPB1*23:01 DQA1*05:01 DQA1*01:02 DQB1*02:01 DQB1*06:02 DRA*01:02 DRA*01:02 DRB1*03:01 DRB1*15:01
C006 A*03:01 A*01:01 B*07:02 B*08:01 C*07:01 C*07:02 DPA1*01:03 DPA1*01:03 DPB1*57:01 DPB1*271:01 DQA1*01:02 DQA1*01:01 DQB1*06:04 DQB1*05:01 DRA*01:01 DRA*01:02 DRB1*13:02 DRB1*01:02
C007 A*01:01 A*02:01 B*15:01 B*08:01 C*07:01 C*03:03 DPA1*02:01 DPA1*01:03 DPB1*04:02 DPB1*13:01 DQA1*05:01 DQA1*03:01 DQB1*02:01 DQB1*03:02 DRA*01:01 DRA*01:02 DRB1*03:01 DRB1*04:01
C008 A*11:01 A*02:01 B*35:01 B*27:05 C*03:04 C*04:01 DPA1*01:03 DPA1*01:03 DPB1*04:01 DPB1*04:01 DQA1*03:03 DQA1*01:02 DQB1*06:04 DQB1*03:01 DRA*01:02 DRA*01:01 DRB1*13:02 DRB1*04:01
C009 A*23:01 A*01:01 B*13:02 B*18:01 C*07:01 C*07:02 DPA1*02:01 DPA1*01:03 DPB1*01:01 DPB1*04:02 DQA1*05:05 DQA1*03:01 DQB1*03:01 DQB1*03:02 DRA*01:01 DRA*01:01 DRB1*11:04 DRB1*04:03
C010 A*31:01 A*02:06 B*15:01 B*56:01 C*04:01 C*03:03 DPA1*02:02 DPA1*02:02 DPB1*05:01 DPB1*05:01 DQA1*03:02 DQA1*03:02 DQB1*03:03 DQB1*03:96 DRA*01:01 DRA*01:01 DRB1*09:01 DRB1*09:01
C011 A*02:01 A*01:01 B*07:02 B*13:02 C*06:02 C*07:02 DPA1*02:01 DPA1*02:01 DPB1*17:01 DPB1*10:01 DQA1*02:01 DQA1*01:01 DQB1*02:02 DQB1*05:01 DRA*01:01 DRA*01:01 DRB1*07:01 DRB1*01:01
C012 A*02:01 A*02:01 B*15:01 B*27:02 C*02:02 C*03:03 DPA1*01:03 DPA1*01:03 DPB1*04:01 DPB1*04:01 DQA1*01:02 DQA1*03:01 DQB1*03:02 DQB1*05:02 DRA*01:01 DRA*01:01 DRB1*04:04 DRB1*16:01
C013 A*02:05 A*01:01 B*49:01 B*57:01 C*07:01 C*07:01 DPA1*01:03 DPA1*01:03 DPB1*04:01 DPB1*04:01 DQA1*02:01 DQA1*03:03 DQB1*03:02 DQB1*03:03 DRA*01:01 DRA*01:01 DRB1*04:05 DRB1*07:01
C014 A*30:02 A*01:01 B*37:01 B*27:05 C*06:02 C*02:02 DPA1*02:01 DPA1*01:03 DPB1*02:01 DPB1*10:01 DQA1*03:01 DQA1*03:03 DQB1*04:02 DQB1*03:02 DRA*01:01 DRA*01:01 DRB1*04:03 DRB1*04:04
C015 A*02:642 A*03:01 B*07:02 B*07:02 C*07:02 C*07:02 DPA1*01:03 DPA1*01:03 DPB1*04:01 DPB1*16:01 DQA1*01:02 DQA1*01:02 DQB1*06:02 DQB1*06:02 DRA*01:02 DRA*01:02 DRB1*15:01 DRB1*15:01
C016 A*01:01 A*68:01 B*42:01 B*08:156 C*07:02 C*07:01 DPA1*01:03 DPA1*01:03 DPB1*04:01 DPB1*02:01 DQA1*05:01 DQA1*01:02 DQB1*02:01 DQB1*06:02 DRA*01:02 DRA*01:02 DRB1*03:01 DRB1*15:01
C017 A*03:01 A*11:01 B*08:01 B*18:01 C*05:01 C*07:01 DPA1*02:01 DPA1*01:03 DPB1*04:01 DPB1*01:01 DQA1*05:01 DQA1*05:01 DQB1*02:01 DQB1*02:01 DRA*01:01 DRA*01:02 DRB1*03:01 DRB1*03:01
C018 A*11:01 A*02:01 B*15:01 B*37:01 C*06:02 C*04:01 DPA1*01:03 DPA1*02:02 DPB1*04:01 DPB1*05:01 DQA1*01:03 DQA1*01:05 DQB1*06:01 DQB1*05:01 DRA*01:01 DRA*01:02 DRB1*10:01 DRB1*08:03
C019 A*02:10 A*31:01 B*54:01 B*40:06 C*08:01 C*01:02 DPA1*02:02 DPA1*01:03 DPB1*04:02 DPB1*05:01 DQA1*03:02 DQA1*03:01 DQB1*04:01 DQB1*03:96 DRA*01:01 DRA*01:01 DRB1*09:01 DRB1*04:05
C020 A*01:01 A*29:02 B*44:03 B*08:01 C*07:01 C*16:01 DPA1*02:01 DPA1*02:02 DPB1*01:01 DPB1*10:01 DQA1*02:01 DQA1*05:01 DQB1*02:01 DQB1*02:02 DRA*01:02 DRA*01:01 DRB1*07:01 DRB1*03:01

Next, we want to check our HLA allele frequencies, and compare them to known frequencies from major populations. Here, we only include alleles with an allele frequency of 5% or higher in our study cohort. By default, MiDAS will output comparisons including the following populations, based on published data from allelefrequencies.net:

MiDAS comes with some pre-defined reference populations, but it is possible to customize these comparisons (see documentation).

freq_HLA <- getHlaFrequencies(hla_calls = dat_HLA, compare = TRUE) %>%
  filter(Freq > 0.01)
HLA frequencies, compared to published references
allele Counts Freq USA NMDP African American pop 2 USA NMDP Chinese USA NMDP European Caucasian USA NMDP Hispanic South or Central American USA NMDP Japanese USA NMDP North American Amerindian USA NMDP South Asian Indian
A*01:01 236 0.1180 0.0467 0.0145 0.1646 0.0726 0.0100 0.1202 0.1545
A*02:01 486 0.2430 0.1235 0.0946 0.2755 0.2095 0.1480 0.2776 0.0492
A*02:06 22 0.0110 0.0007 0.0349 0.0018 0.0198 0.0748 0.0275 0.0175
A*03:01 199 0.0995 0.0839 0.0140 0.1399 0.0738 0.0090 0.1044 0.0636
A*11:01 114 0.0570 0.0142 0.2752 0.0609 0.0456 0.0874 0.0488 0.1396
A*23:01 47 0.0235 0.1099 0.0021 0.0197 0.0368 0.0011 0.0181 0.0066

Let’s assume our cohort is of predominantly European ancestry. We can plot the following comparison to see if allele frequencies in our data are distributed as expected, for example by visualizing common HLA-A allele frequencies in comparison with European, Chinese, and African American reference populations:

freq_HLA_long <- tidyr::gather(
  data = freq_HLA,
  key = "population",
  value = "freq",
  "Freq",
  "USA NMDP European Caucasian",
  "USA NMDP Chinese",
  "USA NMDP African American pop 2",
  factor_key = TRUE
) %>% 
  filter(grepl("^A", allele))

plot_HLAfreq <-
  ggplot2::ggplot(data = freq_HLA_long, ggplot2::aes(x = allele, y = freq, fill = population)) +
  ggplot2::geom_bar(
    stat = "identity",
    position = ggplot2::position_dodge(0.7),
    width = 0.7,
    colour = "black"
  ) +
  ggplot2::coord_flip() +
  ggplot2::scale_y_continuous(labels = formattable::percent)

plot_HLAfreq

HLA association analysis

Are classical HLA alleles associated with disease status?

The following function prepares our data for analysis, combining HLA and phenotypic data into one object. Here, we are interested in analyzing our data on the level of classical HLA alleles.

We can now test our HLA data for deviations from Hardy-Weinberg-Equilibrium (HWE) to filter out alleles that strongly deviate from HWE expectations (imputation or genotyping errors, …). Here, let’s remove alleles with a HWE p-value below 0.05 divided by the number of alleles tested / present in our data (N=447). We can create a filtered MiDAS object right away (as.MiDAS = TRUE), as done in this example, or output actual HWE test results.

Now, we define our statistical model and run the analysis. Since we want to test for association with disease status, we use a logistic regression approach. The term is necessary as placeholder for the tested HLA alleles and needs to be included. It will become handy when for example defining more complex interaction models.

In the runMiDAS function, we then refer to this model, choose our analysis type and define a inheritance model. Here we use dominant model, meaning that individuals will be defined as non-carriers (0) vs. carriers (1) for a given allele. Alternatively, it is also possible to choose recessive (0 = non-carrier or heterozygous carrier, 1 = homozygous carrier), overdominant (assuming heterozygous (dis)advantage: 0 = non-carrier or homozygous carrier, 1 = heterozygous carrier), or additive (N of alleles) inheritance models. Moreover, we define allele inclusion criteria, such that we only consider alleles frequencies above 2% and below 98%. We also apply the Bonferroni method to not only get nominal P values, but also such adjusted for multiple testing. For alternative multiple testing correction methods, as well as the option to choose a custom number of tests, please refer to the documentation. exponentiate = TRUE means that the effect estimate will already be shown as odds ratio, since we use a logistic regression model.

MiDAS analysis results
allele p.value p.adjusted estimate std.error conf.low conf.high statistic Ntotal Ntotal [%] N(disease=0) N(disease=0) [%] N(disease=1) N(disease=1) [%]
DQB1*06:02 2.600e-06 2.290e-04 2.1719 0.1651 1.5764 3.0138 4.698e+00 198 9.90% 69 6.90% 129 12.90%
DRB1*15:01 7.300e-06 6.386e-04 2.0649 0.1617 1.5081 2.8450 4.484e+00 206 10.30% 74 7.40% 132 13.20%
B*57:01 3.080e-05 2.682e-03 2.9931 0.2631 1.8157 5.1189 4.167e+00 79 3.95% 21 2.10% 58 5.80%
C*07:02 1.859e-04 1.618e-02 1.8227 0.1606 1.3334 2.5044 3.737e+00 204 10.20% 78 7.80% 126 12.60%
B*18:01 6.012e-04 5.230e-02 0.5014 0.2012 0.3357 0.7400 -3.431e+00 122 6.10% 79 7.90% 43 4.30%
DRA*01:02 2.701e-03 2.350e-01 1.4703 0.1285 1.1435 1.8926 3.000e+00 573 28.65% 263 26.30% 310 31.00%
DQB1*02:02 4.569e-03 3.975e-01 0.6257 0.1654 0.4513 0.8635 -2.836e+00 185 9.25% 110 11.00% 75 7.50%
B*51:01 4.863e-03 4.231e-01 1.7271 0.1941 1.1849 2.5396 2.816e+00 128 6.40% 49 4.90% 79 7.90%
DRA*01:01 1.073e-02 9.331e-01 0.6210 0.1867 0.4289 0.8930 -2.552e+00 862 43.10% 445 44.50% 417 41.70%
B*44:02 2.068e-02 1.000e+00 1.4963 0.1742 1.0655 2.1111 2.314e+00 161 8.05% 67 6.70% 94 9.40%
C*08:02 2.076e-02 1.000e+00 1.8030 0.2549 1.1023 3.0076 2.312e+00 71 3.55% 26 2.60% 45 4.50%
DQA1*04:01 2.235e-02 1.000e+00 1.9179 0.2851 1.1089 3.4120 2.284e+00 57 2.85% 20 2.00% 37 3.70%
DQB1*05:02 3.106e-02 1.000e+00 0.5645 0.2652 0.3313 0.9417 -2.156e+00 65 3.25% 41 4.10% 24 2.40%
DRB1*11:01 3.188e-02 1.000e+00 0.6651 0.1900 0.4565 0.9630 -2.146e+00 131 6.55% 77 7.70% 54 5.40%
DRB1*07:01 3.518e-02 1.000e+00 0.7266 0.1517 0.5390 0.9772 -2.106e+00 228 11.40% 128 12.80% 100 10.00%
A*31:01 6.065e-02 1.000e+00 1.6857 0.2784 0.9846 2.9492 1.876e+00 58 2.90% 22 2.20% 36 3.60%
B*08:01 8.299e-02 1.000e+00 0.7130 0.1951 0.4847 1.0431 -1.734e+00 122 6.10% 70 7.00% 52 5.20%
DQA1*01:02 9.218e-02 1.000e+00 1.2555 0.1351 0.9637 1.6371 1.684e+00 327 16.35% 151 15.10% 176 17.60%
B*07:02 1.037e-01 1.000e+00 1.3456 0.1825 0.9424 1.9292 1.627e+00 142 7.10% 62 6.20% 80 8.00%
B*35:01 1.043e-01 1.000e+00 0.6811 0.2365 0.4254 1.0787 -1.624e+00 80 4.00% 47 4.70% 33 3.30%
DQB1*04:02 1.069e-01 1.000e+00 1.5610 0.2762 0.9141 2.7147 1.612e+00 58 2.90% 23 2.30% 35 3.50%
DRB1*03:01 1.102e-01 1.000e+00 0.7737 0.1606 0.5639 1.0592 -1.597e+00 194 9.70% 107 10.70% 87 8.70%
DQA1*02:01 1.177e-01 1.000e+00 0.7913 0.1496 0.5895 1.0604 -1.565e+00 235 11.75% 128 12.80% 107 10.70%
C*03:03 1.291e-01 1.000e+00 0.6974 0.2374 0.4349 1.1070 -1.518e+00 79 3.95% 46 4.60% 33 3.30%
C*07:01 1.309e-01 1.000e+00 0.7952 0.1517 0.5901 1.0700 -1.511e+00 226 11.30% 123 12.30% 103 10.30%
DQB1*02:01 1.314e-01 1.000e+00 0.7860 0.1596 0.5740 1.0739 -1.509e+00 197 9.85% 108 10.80% 89 8.90%
DPB1*104:01 1.341e-01 1.000e+00 0.6295 0.3089 0.3380 1.1441 -1.498e+00 46 2.30% 28 2.80% 18 1.80%
DQA1*01:01 1.613e-01 1.000e+00 1.2610 0.1656 0.9122 1.7473 1.401e+00 179 8.95% 81 8.10% 98 9.80%
DQB1*03:01 1.699e-01 1.000e+00 0.8355 0.1309 0.6461 1.0797 -1.373e+00 373 18.65% 197 19.70% 176 17.60%
A*02:01 1.796e-01 1.000e+00 1.1874 0.1280 0.9241 1.5265 1.342e+00 427 21.35% 203 20.30% 224 22.40%
DQA1*05:05 2.041e-01 1.000e+00 0.8356 0.1414 0.6330 1.1022 -1.270e+00 278 13.90% 148 14.80% 130 13.00%
A*11:01 2.223e-01 1.000e+00 0.7788 0.2049 0.5194 1.1620 -1.220e+00 108 5.40% 60 6.00% 48 4.80%
DPA1*01:03 2.340e-01 1.000e+00 1.3315 0.2406 0.8330 2.1466 1.190e+00 924 46.20% 457 45.70% 467 46.70%
DQB1*03:03 2.340e-01 1.000e+00 1.3315 0.2406 0.8330 2.1466 1.190e+00 76 3.80% 33 3.30% 43 4.30%
DQA1*05:01 2.369e-01 1.000e+00 0.8294 0.1582 0.6076 1.1303 -1.183e+00 201 10.05% 108 10.80% 93 9.30%
A*25:01 2.563e-01 1.000e+00 0.7211 0.2881 0.4056 1.2630 -1.135e+00 52 2.60% 30 3.00% 22 2.20%
B*27:05 2.728e-01 1.000e+00 0.7374 0.2778 0.4239 1.2666 -1.097e+00 56 2.80% 32 3.20% 24 2.40%
DPB1*04:02 2.921e-01 1.000e+00 0.8426 0.1625 0.6120 1.1582 -1.054e+00 187 9.35% 100 10.00% 87 8.70%
A*03:01 2.960e-01 1.000e+00 1.1835 0.1612 0.8632 1.6250 1.045e+00 191 9.55% 89 8.90% 102 10.20%
DPB1*17:01 2.973e-01 1.000e+00 0.7299 0.3020 0.3991 1.3139 -1.042e+00 47 2.35% 27 2.70% 20 2.00%
C*03:04 2.978e-01 1.000e+00 1.2747 0.2331 0.8086 2.0229 1.041e+00 81 4.05% 36 3.60% 45 4.50%
DQA1*03:01 2.986e-01 1.000e+00 1.1981 0.1738 0.8526 1.6870 1.040e+00 158 7.90% 73 7.30% 85 8.50%
DPB1*13:01 3.067e-01 1.000e+00 0.7390 0.2959 0.4093 1.3149 -1.022e+00 49 2.45% 28 2.80% 21 2.10%
DQA1*01:03 3.208e-01 1.000e+00 1.1970 0.1811 0.8398 1.7102 9.928e-01 143 7.15% 66 6.60% 77 7.70%
DRB1*04:01 3.369e-01 1.000e+00 1.2288 0.2145 0.8079 1.8775 9.604e-01 97 4.85% 44 4.40% 53 5.30%
DQB1*05:03 3.760e-01 1.000e+00 0.7680 0.2981 0.4239 1.3736 -8.854e-01 48 2.40% 27 2.70% 21 2.10%
DQA1*01:04 3.938e-01 1.000e+00 0.7831 0.2867 0.4426 1.3704 -8.528e-01 52 2.60% 29 2.90% 23 2.30%
A*32:01 4.177e-01 1.000e+00 0.8022 0.2720 0.4674 1.3647 -8.104e-01 58 2.90% 32 3.20% 26 2.60%
A*23:01 4.558e-01 1.000e+00 0.7993 0.3005 0.4394 1.4375 -7.457e-01 47 2.35% 26 2.60% 21 2.10%
DQB1*06:01 4.558e-01 1.000e+00 1.2512 0.3005 0.6956 2.2759 7.457e-01 47 2.35% 21 2.10% 26 2.60%
B*40:01 4.958e-01 1.000e+00 1.2050 0.2738 0.7053 2.0743 6.812e-01 57 2.85% 26 2.60% 31 3.10%
C*12:03 5.147e-01 1.000e+00 0.8856 0.1865 0.6135 1.2760 -6.516e-01 133 6.65% 70 7.00% 63 6.30%
DRB1*13:01 5.273e-01 1.000e+00 1.1428 0.2112 0.7557 1.7330 6.321e-01 100 5.00% 47 4.70% 53 5.30%
A*29:02 5.379e-01 1.000e+00 0.8264 0.3096 0.4463 1.5145 -6.159e-01 44 2.20% 24 2.40% 20 2.00%
DRB1*01:01 5.622e-01 1.000e+00 0.8939 0.1934 0.6108 1.3059 -5.795e-01 122 6.10% 64 6.40% 58 5.80%
A*68:01 5.692e-01 1.000e+00 0.8499 0.2857 0.4823 1.4870 -5.692e-01 52 2.60% 28 2.80% 24 2.40%
B*44:03 5.709e-01 1.000e+00 0.8792 0.2271 0.5616 1.3720 -5.667e-01 85 4.25% 45 4.50% 40 4.00%
DRB1*13:02 5.750e-01 1.000e+00 1.1343 0.2247 0.7303 1.7673 5.607e-01 87 4.35% 41 4.10% 46 4.60%
DRB1*11:04 5.789e-01 1.000e+00 0.8839 0.2224 0.5699 1.3666 -5.550e-01 89 4.45% 47 4.70% 42 4.20%
DQB1*03:02 6.084e-01 1.000e+00 1.0915 0.1709 0.7808 1.5273 5.123e-01 164 8.20% 79 7.90% 85 8.50%
B*13:02 6.202e-01 1.000e+00 0.8843 0.2483 0.5415 1.4384 -4.955e-01 70 3.50% 37 3.70% 33 3.30%
DPA1*02:01 6.217e-01 1.000e+00 0.9328 0.1410 0.7072 1.2298 -4.935e-01 279 13.95% 143 14.30% 136 13.60%
B*15:01 6.412e-01 1.000e+00 0.8969 0.2334 0.5659 1.4173 -4.661e-01 80 4.00% 42 4.20% 38 3.80%
DRB1*04:04 6.542e-01 1.000e+00 0.8745 0.2994 0.4829 1.5731 -4.480e-01 47 2.35% 25 2.50% 22 2.20%
B*38:01 6.665e-01 1.000e+00 0.8833 0.2880 0.4993 1.5537 -4.310e-01 51 2.55% 27 2.70% 24 2.40%
DQB1*06:04 7.083e-01 1.000e+00 1.0980 0.2498 0.6727 1.7976 3.742e-01 69 3.45% 33 3.30% 36 3.60%
DPB1*03:01 7.087e-01 1.000e+00 0.9325 0.1869 0.6457 1.3454 -3.736e-01 132 6.60% 68 6.80% 64 6.40%
DPB1*02:01 7.233e-01 1.000e+00 0.9511 0.1417 0.7203 1.2555 -3.541e-01 275 13.75% 140 14.00% 135 13.50%
C*12:02 7.526e-01 1.000e+00 0.9053 0.3156 0.4842 1.6832 -3.152e-01 42 2.10% 22 2.20% 20 2.00%
A*24:02 7.586e-01 1.000e+00 1.0484 0.1537 0.7756 1.4176 3.074e-01 216 10.80% 106 10.60% 110 11.00%
C*05:01 7.899e-01 1.000e+00 0.9538 0.1776 0.6728 1.3513 -2.664e-01 149 7.45% 76 7.60% 73 7.30%
DPB1*01:01 8.067e-01 1.000e+00 0.9419 0.2448 0.5814 1.5231 -2.446e-01 72 3.60% 37 3.70% 35 3.50%
C*02:02 8.215e-01 1.000e+00 1.0522 0.2256 0.6756 1.6408 2.256e-01 86 4.30% 42 4.20% 44 4.40%
C*01:02 8.316e-01 1.000e+00 1.0463 0.2128 0.6891 1.5902 2.127e-01 98 4.90% 48 4.80% 50 5.00%
DQB1*06:03 8.385e-01 1.000e+00 1.0424 0.2038 0.6987 1.5563 2.038e-01 108 5.40% 53 5.30% 55 5.50%
DQB1*05:01 8.724e-01 1.000e+00 1.0261 0.1606 0.7489 1.4063 1.606e-01 192 9.60% 95 9.50% 97 9.70%
A*30:02 8.761e-01 1.000e+00 0.9526 0.3119 0.5141 1.7603 -1.559e-01 43 2.15% 22 2.20% 21 2.10%
DRB1*04:05 8.812e-01 1.000e+00 1.0457 0.2989 0.5806 1.8875 1.494e-01 47 2.35% 23 2.30% 24 2.40%
C*15:02 8.812e-01 1.000e+00 0.9563 0.2989 0.5298 1.7224 -1.494e-01 47 2.35% 24 2.40% 23 2.30%
C*16:01 8.835e-01 1.000e+00 0.9580 0.2930 0.5370 1.7054 -1.465e-01 49 2.45% 25 2.50% 24 2.40%
DQA1*03:03 9.248e-01 1.000e+00 0.9824 0.1887 0.6781 1.4227 -9.434e-02 129 6.45% 65 6.50% 64 6.40%
C*06:02 9.348e-01 1.000e+00 0.9867 0.1636 0.7158 1.3601 -8.178e-02 183 9.15% 92 9.20% 91 9.10%
C*04:01 9.396e-01 1.000e+00 1.0115 0.1515 0.7516 1.3615 7.573e-02 225 11.25% 112 11.20% 113 11.30%
B*49:01 1.000e+00 1.000e+00 1.0000 0.3153 0.5367 1.8633 0.000e+00 42 2.10% 21 2.10% 21 2.10%
A*26:01 1.000e+00 1.000e+00 1.0000 0.2232 0.6447 1.5510 0.000e+00 88 4.40% 44 4.40% 44 4.40%
B*52:01 1.000e+00 1.000e+00 1.0000 0.3153 0.5367 1.8633 0.000e+00 42 2.10% 21 2.10% 21 2.10%
A*01:01 1.000e+00 1.000e+00 1.0000 0.1522 0.7419 1.3479 0.000e+00 222 11.10% 111 11.10% 111 11.10%

Three HLA alleles show significant association with the disease after multiple testing adjustment. Due to the complex linkage disequilibrium structure in the MHC, it is likely that associations are not statistically independent. The two alleles HLA-DRB1*15:01 and HLA-DQB1*06:02 are a common class II haplotype. We can therefore test if there are associations that are statistically independent of our top-associated allele, by setting the conditional flag to TRUE. MiDAS will now perform stepwise conditional testing until the top associated allele does not reach a defined significance threshold (here th = 0.05, based on adjusted p value).

MiDAS analysis results
allele p.value p.adjusted estimate std.error conf.low conf.high statistic covariates Ntotal Ntotal [%] N(disease=0) N(disease=0) [%] N(disease=1) N(disease=1) [%]
DQB1*06:02 2.60e-06 2.290e-04 2.172 0.1651 1.576 3.014 4.698 198 9.90% 69 6.90% 129 12.90%
B*57:01 2.64e-05 2.269e-03 3.049 0.2653 1.841 5.235 4.203 DQB1*06:02 79 3.95% 21 2.10% 58 5.80%

The results for conditional testing are displayed in a way that for each step the top associated allele is shown, along with a list of alleles conditioned on.

As we can see, HLA-DRB1*15:01 was not independently associated with the disease when correcting for our top-associated allele HLA-DQB1*06:02. However, HLA-B*57:01 can be considered an independent association signal.

HLA association fine-mapping on amino acid level

Next, we want to find out what are the strongest associated amino acid positions, corresponding to our allele-level associations. This can help fine-mapping the associated variants to e.g. the peptide binding region or other functionally distinct parts of the protein. We thus prepare a MiDAS object with experiment type “hla_aa”, which includes the inference of amino acid variation from allele calls.

HLA_AA <- prepareMiDAS(
  hla_calls = dat_HLA,
  colData = dat_pheno,
  experiment = "hla_aa"
)

Amino acid data will be stored in a MiDAS object, but we can extract it to a data frame and select a couple of variables to display how this looks like:

dat_HLA_AA <- HLA_AA[["hla_aa"]] %>% 
  assay() %>% 
  t() %>% 
  as.data.frame() %>% 
  select(starts_with("B_97_")) %>% 
  head()
HLA amino acid data as inferred by MiDAS
B_97_S B_97_R B_97_T B_97_N B_97_V B_97_W
C001 1 1 0 0 0 0
C002 0 2 0 0 0 0
C003 0 1 1 0 0 0
C004 1 1 0 0 0 0
C005 1 1 0 0 0 0
C006 2 0 0 0 0 0

Now, we run the association test based on amino acid variation. To first identify the most relevant associated amino acid positions, we run a likelihood ratio (omnibus) test, which groups all residues at each amino acid position.

HLA_AA_model <- glm(disease ~ term, data = HLA_AA, family = binomial())
HLA_AA_omnibus_results <- runMiDAS(
  HLA_AA_model,
  experiment = "hla_aa",
  inheritance_model = "dominant",
  conditional = FALSE,
  omnibus = TRUE,
  lower_frequency_cutoff = 0.02,
  upper_frequency_cutoff = 0.98,
  correction = "bonferroni"
)

kableResults(HLA_AA_omnibus_results)
MiDAS analysis results
aa_pos residues df statistic p.value p.adjusted
B_97 S, R, T, N, V, W 6 3.532e+01 3.700e-06 1.573e-03
DQB1_9 F, Y, * 3 2.569e+01 1.110e-05 4.682e-03
DQB1_185 *, T 2 1.881e+01 8.240e-05 3.476e-02
DQB1_142 *, V 2 1.788e+01 1.313e-04 5.542e-02
DQB1_177 *, H 2 1.788e+01 1.313e-04 5.542e-02
DQB1_167 *, R 2 1.729e+01 1.759e-04 7.422e-02
DQB1_116 *, V 2 1.695e+01 2.081e-04 8.782e-02
DQB1_125 *, A 2 1.695e+01 2.081e-04 8.782e-02
B_81 L, A 2 1.689e+01 2.154e-04 9.092e-02
DQB1_130 *, R 2 1.663e+01 2.451e-04 1.034e-01
DQB1_126 *, Q 2 1.636e+01 2.807e-04 1.185e-01
B_80 N, T, I 3 1.866e+01 3.209e-04 1.354e-01
B_82 R, L 2 1.582e+01 3.674e-04 1.550e-01
B_83 G, R 2 1.582e+01 3.674e-04 1.550e-01
DQB1_140 *, A, T 3 1.836e+01 3.711e-04 1.566e-01
DQB1_182 *, S, N 3 1.836e+01 3.711e-04 1.566e-01
DQB1_135 *, G, D 3 1.815e+01 4.087e-04 1.725e-01
B_62 R, G 2 1.511e+01 5.232e-04 2.208e-01
DRB1_71 A, R, E, K 4 1.984e+01 5.362e-04 2.263e-01
B_275 *, E 2 1.476e+01 6.231e-04 2.630e-01
B_295 *, . 2 1.476e+01 6.231e-04 2.630e-01
B_296 *, A 2 1.476e+01 6.231e-04 2.630e-01
B_297 *, V 2 1.476e+01 6.231e-04 2.630e-01
B_299 *, V 2 1.476e+01 6.231e-04 2.630e-01
B_300 *, I 2 1.476e+01 6.231e-04 2.630e-01
B_301 *, G 2 1.476e+01 6.231e-04 2.630e-01
B_77 S, N, D 3 1.717e+01 6.530e-04 2.756e-01
B_65 Q, R 2 1.385e+01 9.828e-04 4.147e-01
B_70 Q, N, K, S 4 1.849e+01 9.909e-04 4.181e-01
B_239 *, R 2 1.362e+01 1.105e-03 4.664e-01
B_253 *, E 2 1.362e+01 1.105e-03 4.664e-01
B_267 *, P 2 1.362e+01 1.105e-03 4.664e-01
B_268 *, K 2 1.362e+01 1.105e-03 4.664e-01
B_270 *, L 2 1.362e+01 1.105e-03 4.664e-01
B_66 I, N 2 1.330e+01 1.294e-03 5.460e-01
B_194 *, I 2 1.312e+01 1.415e-03 5.970e-01
B_103 V, L 2 1.296e+01 1.537e-03 6.484e-01
B_30 D, G 2 1.283e+01 1.637e-03 6.907e-01
B_-16 *, V 2 1.255e+01 1.886e-03 7.960e-01
B_339 *, A 2 1.242e+01 2.010e-03 8.481e-01
B_199 *, A 2 1.215e+01 2.298e-03 9.696e-01
B_211 *, A 2 1.207e+01 2.395e-03 1.000e+00
DQB1_55 R, L, P 3 1.420e+01 2.647e-03 1.000e+00
C_99 Y, C, F, S 4 1.624e+01 2.714e-03 1.000e+00
DRA_217 L, V 2 1.173e+01 2.830e-03 1.000e+00
B_9 Y, H, D 3 1.370e+01 3.340e-03 1.000e+00
B_69 A, T 2 1.122e+01 3.661e-03 1.000e+00
DQB1_38 A, V 2 1.081e+01 4.489e-03 1.000e+00
DQB1_77 T, R, * 3 1.287e+01 4.925e-03 1.000e+00
B_67 Y, S, F, C, M 5 1.666e+01 5.181e-03 1.000e+00
B_71 A, T 2 1.044e+01 5.405e-03 1.000e+00
DQB1_87 F, L, *, Y 4 1.424e+01 6.568e-03 1.000e+00
DQB1_30 Y, S, H 3 1.148e+01 9.382e-03 1.000e+00
B_-23 *, R 2 9.202e+00 1.004e-02 1.000e+00
B_-21 *, T 2 9.202e+00 1.004e-02 1.000e+00
DQB1_28 T, S 2 9.008e+00 1.106e-02 1.000e+00
DQB1_46 V, E 2 9.008e+00 1.106e-02 1.000e+00
DQB1_47 Y, F 2 9.008e+00 1.106e-02 1.000e+00
DQB1_52 P, L 2 9.008e+00 1.106e-02 1.000e+00
B_-10 *, G 2 8.724e+00 1.275e-02 1.000e+00
B_-9 *, A 2 8.473e+00 1.446e-02 1.000e+00
DQB1_74 E, A, S, * 4 1.222e+01 1.579e-02 1.000e+00
B_163 E, L, T 3 1.031e+01 1.610e-02 1.000e+00
B_282 * 1 5.561e+00 1.836e-02 1.000e+00
B_306 * 1 5.561e+00 1.836e-02 1.000e+00
B_326 * 1 5.561e+00 1.836e-02 1.000e+00
DQA1_107 I, *, T 3 9.899e+00 1.945e-02 1.000e+00
DQA1_156 L, *, F 3 9.899e+00 1.945e-02 1.000e+00
DQA1_161 E, *, D 3 9.899e+00 1.945e-02 1.000e+00
DQA1_163 S, *, I 3 9.899e+00 1.945e-02 1.000e+00
DQB1_71 T, K, D, A, * 5 1.324e+01 2.121e-02 1.000e+00
DQB1_57 D, A, V, S 4 1.140e+01 2.237e-02 1.000e+00
DQB1_37 Y, I, D 3 9.592e+00 2.237e-02 1.000e+00
DRB1_25 R, Q 2 7.501e+00 2.350e-02 1.000e+00
DRB1_14 E, K 2 7.447e+00 2.415e-02 1.000e+00
DQA1_175 K, *, Q, E 4 1.122e+01 2.421e-02 1.000e+00
DRB1_13 R, Y, S, F, H, G 6 1.442e+01 2.532e-02 1.000e+00
DQA1_75 S, I 2 7.203e+00 2.729e-02 1.000e+00
DQB1_75 L, V, * 3 9.131e+00 2.760e-02 1.000e+00
DPB1_85 G, E 2 7.165e+00 2.781e-02 1.000e+00
DPB1_86 P, A 2 7.165e+00 2.781e-02 1.000e+00
DPB1_87 M, V 2 7.165e+00 2.781e-02 1.000e+00
DQA1_-16 M, *, L 3 8.964e+00 2.977e-02 1.000e+00
DRB1_11 P, G, S, V, L 5 1.229e+01 3.107e-02 1.000e+00
B_-8 *, V 2 6.935e+00 3.120e-02 1.000e+00
DQA1_69 L, A, T 3 8.813e+00 3.189e-02 1.000e+00
DPB1_56 E, A 2 6.804e+00 3.331e-02 1.000e+00
B_-11 *, W 2 6.510e+00 3.858e-02 1.000e+00
DRB1_73 A, G 2 6.484e+00 3.909e-02 1.000e+00
DQA1_56 ., G, R 3 8.283e+00 4.051e-02 1.000e+00
DQA1_76 L, M, V 3 8.283e+00 4.051e-02 1.000e+00
B_63 N, E 2 6.264e+00 4.363e-02 1.000e+00
B_143 T, S 2 6.242e+00 4.412e-02 1.000e+00
B_147 W, L 2 6.242e+00 4.412e-02 1.000e+00
DQB1_53 Q, L 2 6.167e+00 4.581e-02 1.000e+00
DRB1_70 Q, D, R 3 7.952e+00 4.701e-02 1.000e+00
B_178 K, T 2 6.097e+00 4.744e-02 1.000e+00
DQA1_129 H, *, Q 3 7.675e+00 5.323e-02 1.000e+00
DRB1_78 Y, V 2 5.806e+00 5.486e-02 1.000e+00
B_45 E, K, M, T 4 9.174e+00 5.689e-02 1.000e+00
B_152 E, V 2 5.717e+00 5.736e-02 1.000e+00
A_156 W, L, Q, R 4 9.095e+00 5.876e-02 1.000e+00
DPB1_84 G, D 2 5.518e+00 6.335e-02 1.000e+00
DQA1_47 C, K, R, Q 4 8.866e+00 6.454e-02 1.000e+00
B_167 W, S 2 5.462e+00 6.514e-02 1.000e+00
DPB1_69 E, K, R 3 7.206e+00 6.560e-02 1.000e+00
DRB1_96 *, H 2 5.296e+00 7.080e-02 1.000e+00
DRB1_98 *, K 2 5.296e+00 7.080e-02 1.000e+00
DRB1_104 *, S 2 5.296e+00 7.080e-02 1.000e+00
DRB1_120 *, S 2 5.296e+00 7.080e-02 1.000e+00
DRB1_149 *, H 2 5.296e+00 7.080e-02 1.000e+00
DRB1_180 *, V 2 5.296e+00 7.080e-02 1.000e+00
DRB1_58 A, E 2 5.167e+00 7.550e-02 1.000e+00
DQA1_54 F, L 2 4.945e+00 8.436e-02 1.000e+00
DQB1_56 P, L 2 4.874e+00 8.743e-02 1.000e+00
DRB1_74 A, Q, E, R, L 5 9.459e+00 9.208e-02 1.000e+00
B_116 Y, L, S, F, D 5 9.438e+00 9.281e-02 1.000e+00
DRB1_77 T, N 2 4.705e+00 9.511e-02 1.000e+00
DQB1_66 E, D 2 4.696e+00 9.554e-02 1.000e+00
DQB1_67 V, I 2 4.696e+00 9.554e-02 1.000e+00
DQB1_23 R 1 2.777e+00 9.565e-02 1.000e+00
B_46 E, A 2 4.545e+00 1.030e-01 1.000e+00
DQA1_18 S, F 2 4.503e+00 1.052e-01 1.000e+00
DQA1_45 V, A 2 4.503e+00 1.052e-01 1.000e+00
DQA1_48 L, W 2 4.503e+00 1.052e-01 1.000e+00
DQA1_55 R, G 2 4.503e+00 1.052e-01 1.000e+00
DQA1_61 F, G 2 4.503e+00 1.052e-01 1.000e+00
DQA1_64 T, R 2 4.503e+00 1.052e-01 1.000e+00
DQA1_66 I, M 2 4.503e+00 1.052e-01 1.000e+00
DQA1_80 S, Y 2 4.503e+00 1.052e-01 1.000e+00
A_163 R, T 2 4.467e+00 1.072e-01 1.000e+00
DQA1_11 Y, C 2 4.452e+00 1.080e-01 1.000e+00
A_90 D, A 2 4.340e+00 1.142e-01 1.000e+00
DPA1_31 M, Q 2 4.297e+00 1.167e-01 1.000e+00
DPA1_50 Q, R 2 4.237e+00 1.202e-01 1.000e+00
A_245 *, A 2 4.225e+00 1.210e-01 1.000e+00
B_171 Y, H 2 4.132e+00 1.267e-01 1.000e+00
DQB1_84 E, Q, * 3 5.693e+00 1.276e-01 1.000e+00
DQB1_85 V, L, * 3 5.693e+00 1.276e-01 1.000e+00
DQB1_89 G, T, * 3 5.693e+00 1.276e-01 1.000e+00
DQB1_90 I, T, * 3 5.693e+00 1.276e-01 1.000e+00
DPB1_36 V, A 2 4.077e+00 1.302e-01 1.000e+00
DQB1_70 G, R, E, * 4 7.108e+00 1.303e-01 1.000e+00
DQA1_130 S, *, A 3 5.630e+00 1.311e-01 1.000e+00
DRB1_37 S, F, N, Y, L 5 8.451e+00 1.331e-01 1.000e+00
DQA1_102 L, * 2 4.014e+00 1.344e-01 1.000e+00
DQA1_138 T, * 2 4.014e+00 1.344e-01 1.000e+00
DQA1_139 S, * 2 4.014e+00 1.344e-01 1.000e+00
DRB1_10 Q, Y 2 3.965e+00 1.377e-01 1.000e+00
DQA1_50 V, L, E 3 5.421e+00 1.435e-01 1.000e+00
DQA1_53 Q, R, K 3 5.421e+00 1.435e-01 1.000e+00
DQA1_52 R, H, S 3 5.388e+00 1.455e-01 1.000e+00
A_105 P, S 2 3.819e+00 1.482e-01 1.000e+00
A_193 *, A 2 3.707e+00 1.567e-01 1.000e+00
A_194 *, V 2 3.707e+00 1.567e-01 1.000e+00
A_207 *, S 2 3.707e+00 1.567e-01 1.000e+00
A_253 *, Q 2 3.707e+00 1.567e-01 1.000e+00
DQA1_-13 T, *, A 3 5.123e+00 1.630e-01 1.000e+00
DPA1_83 T, A 2 3.592e+00 1.660e-01 1.000e+00
A_77 S, D, N 3 4.931e+00 1.770e-01 1.000e+00
DRB1_12 K, T 2 3.424e+00 1.805e-01 1.000e+00
C_66 K, N 2 3.422e+00 1.807e-01 1.000e+00
DRB1_60 Y, S, H 3 4.810e+00 1.863e-01 1.000e+00
DQB1_45 G, E 2 3.314e+00 1.907e-01 1.000e+00
DRB1_30 Y, L, C, H 4 6.083e+00 1.930e-01 1.000e+00
DQA1_40 G, E 2 3.267e+00 1.953e-01 1.000e+00
DQA1_51 L, F 2 3.267e+00 1.953e-01 1.000e+00
C_152 E, A 2 3.198e+00 2.021e-01 1.000e+00
C_77 S, N 2 3.197e+00 2.022e-01 1.000e+00
C_80 N, K 2 3.197e+00 2.022e-01 1.000e+00
DPA1_96 *, P 1 1.600e+00 2.059e-01 1.000e+00
B_113 H, Y 2 3.079e+00 2.145e-01 1.000e+00
A_73 T, I 2 3.047e+00 2.180e-01 1.000e+00
DRB1_28 D, E 2 3.036e+00 2.192e-01 1.000e+00
DQA1_160 A, *, D 3 4.381e+00 2.231e-01 1.000e+00
DPB1_55 D, E, A 3 4.338e+00 2.272e-01 1.000e+00
DRB1_33 N, H 2 2.928e+00 2.313e-01 1.000e+00
DQA1_1 E, * 2 2.925e+00 2.317e-01 1.000e+00
DQA1_218 R, *, Q 3 4.267e+00 2.340e-01 1.000e+00
DQA1_41 R, K 2 2.893e+00 2.354e-01 1.000e+00
A_43 Q 1 1.387e+00 2.389e-01 1.000e+00
B_109 L 1 1.387e+00 2.389e-01 1.000e+00
C_-17 * 1 1.387e+00 2.389e-01 1.000e+00
C_-15 * 1 1.387e+00 2.389e-01 1.000e+00
C_-9 * 1 1.387e+00 2.389e-01 1.000e+00
C_1 * 1 1.387e+00 2.389e-01 1.000e+00
C_175 G 1 1.387e+00 2.389e-01 1.000e+00
C_184 * 1 1.387e+00 2.389e-01 1.000e+00
C_219 * 1 1.387e+00 2.389e-01 1.000e+00
C_253 * 1 1.387e+00 2.389e-01 1.000e+00
C_267 * 1 1.387e+00 2.389e-01 1.000e+00
C_275 * 1 1.387e+00 2.389e-01 1.000e+00
C_285 * 1 1.387e+00 2.389e-01 1.000e+00
C_312 * 1 1.387e+00 2.389e-01 1.000e+00
C_345 * 1 1.387e+00 2.389e-01 1.000e+00
DQB1_86 A, E, *, G 4 5.435e+00 2.455e-01 1.000e+00
DQA1_2 D, *, G 3 4.104e+00 2.505e-01 1.000e+00
C_173 E, K 2 2.722e+00 2.564e-01 1.000e+00
A_219 *, R 2 2.718e+00 2.569e-01 1.000e+00
C_103 L, V 2 2.700e+00 2.592e-01 1.000e+00
C_91 G, R 2 2.649e+00 2.659e-01 1.000e+00
DRB1_16 H, Y 2 2.631e+00 2.684e-01 1.000e+00
DPB1_-14 *, T 2 2.558e+00 2.782e-01 1.000e+00
DRB1_57 D, V, A, S 4 5.068e+00 2.804e-01 1.000e+00
DQA1_26 T, S 2 2.541e+00 2.807e-01 1.000e+00
A_282 *, I 2 2.520e+00 2.836e-01 1.000e+00
A_311 *, K 2 2.520e+00 2.836e-01 1.000e+00
C_330 *, A 2 2.517e+00 2.840e-01 1.000e+00
DRB1_32 Y, H 2 2.498e+00 2.868e-01 1.000e+00
B_94 T, I 2 2.473e+00 2.904e-01 1.000e+00
DQB1_13 G, A, * 3 3.735e+00 2.915e-01 1.000e+00
C_24 A, S 2 2.431e+00 2.966e-01 1.000e+00
C_295 *, A 2 2.316e+00 3.142e-01 1.000e+00
C_311 *, A 2 2.316e+00 3.142e-01 1.000e+00
C_313 *, V 2 2.316e+00 3.142e-01 1.000e+00
C_332 *, S 2 2.316e+00 3.142e-01 1.000e+00
DRB1_47 F, Y 2 2.287e+00 3.187e-01 1.000e+00
A_9 Y, F, S, T 4 4.680e+00 3.218e-01 1.000e+00
B_12 V, M 2 2.186e+00 3.352e-01 1.000e+00
DRB1_67 I, L, F 3 3.360e+00 3.394e-01 1.000e+00
B_11 S, A 2 2.145e+00 3.421e-01 1.000e+00
DQA1_207 V, *, M 3 3.336e+00 3.426e-01 1.000e+00
DPB1_11 G, L 2 2.127e+00 3.452e-01 1.000e+00
C_113 Y, H 2 2.014e+00 3.653e-01 1.000e+00
A_109 L 1 8.077e-01 3.688e-01 1.000e+00
A_231 *, V 2 1.979e+00 3.718e-01 1.000e+00
B_145 R, L 2 1.976e+00 3.723e-01 1.000e+00
DPB1_8 L, V 2 1.958e+00 3.758e-01 1.000e+00
C_309 *, V 2 1.876e+00 3.915e-01 1.000e+00
C_327 *, C 2 1.876e+00 3.915e-01 1.000e+00
C_14 R, W 2 1.834e+00 3.998e-01 1.000e+00
C_49 A, E 2 1.834e+00 3.998e-01 1.000e+00
A_76 E, V, A 3 2.935e+00 4.018e-01 1.000e+00
DPB1_178 *, L 2 1.820e+00 4.026e-01 1.000e+00
B_32 Q, L 2 1.766e+00 4.136e-01 1.000e+00
DQB1_14 M, L, * 3 2.820e+00 4.202e-01 1.000e+00
A_144 Q, K 2 1.719e+00 4.234e-01 1.000e+00
DRB1_9 W, E 2 1.605e+00 4.482e-01 1.000e+00
C_194 *, V 2 1.600e+00 4.493e-01 1.000e+00
C_261 *, V 2 1.600e+00 4.493e-01 1.000e+00
C_273 *, R 2 1.600e+00 4.493e-01 1.000e+00
C_114 D, N 2 1.576e+00 4.547e-01 1.000e+00
C_248 *, V 2 1.560e+00 4.583e-01 1.000e+00
C_211 *, A 2 1.554e+00 4.597e-01 1.000e+00
DPB1_76 M, V, I 3 2.568e+00 4.631e-01 1.000e+00
C_147 W, L 2 1.511e+00 4.697e-01 1.000e+00
C_35 R, Q 2 1.502e+00 4.718e-01 1.000e+00
C_-18 *, R 2 1.478e+00 4.775e-01 1.000e+00
C_-5 *, T 2 1.478e+00 4.775e-01 1.000e+00
C_-1 *, A 2 1.478e+00 4.775e-01 1.000e+00
C_284 *, I 2 1.478e+00 4.775e-01 1.000e+00
C_289 *, A 2 1.478e+00 4.775e-01 1.000e+00
C_291 *, L 2 1.478e+00 4.775e-01 1.000e+00
C_314 *, M 2 1.478e+00 4.775e-01 1.000e+00
C_315 *, C 2 1.478e+00 4.775e-01 1.000e+00
DPB1_33 Q 1 4.988e-01 4.800e-01 1.000e+00
DQA1_208 G, * 2 1.451e+00 4.842e-01 1.000e+00
DRB1_86 V, G 2 1.448e+00 4.847e-01 1.000e+00
DQA1_-7 V, *, M 3 2.441e+00 4.860e-01 1.000e+00
A_297 *, V 2 1.434e+00 4.881e-01 1.000e+00
C_270 *, L 2 1.412e+00 4.937e-01 1.000e+00
A_186 *, K 2 1.397e+00 4.973e-01 1.000e+00
A_-16 *, L 2 1.395e+00 4.979e-01 1.000e+00
A_283 *, P 2 1.395e+00 4.979e-01 1.000e+00
C_21 R, H 2 1.394e+00 4.980e-01 1.000e+00
C_186 *, K 2 1.387e+00 4.997e-01 1.000e+00
C_310 *, V 2 1.387e+00 4.997e-01 1.000e+00
A_184 *, A, P 3 2.353e+00 5.024e-01 1.000e+00
C_73 A, T 2 1.376e+00 5.027e-01 1.000e+00
C_163 T, L, E 3 2.260e+00 5.203e-01 1.000e+00
DPA1_11 A, M 2 1.299e+00 5.224e-01 1.000e+00
A_149 T, A 2 1.254e+00 5.343e-01 1.000e+00
A_56 G, R 2 1.228e+00 5.411e-01 1.000e+00
B_41 A, T 2 1.228e+00 5.412e-01 1.000e+00
A_161 E, D 2 1.204e+00 5.478e-01 1.000e+00
DPB1_9 F, Y, H 3 2.109e+00 5.501e-01 1.000e+00
C_138 T, K 2 1.169e+00 5.573e-01 1.000e+00
A_151 H, R 2 1.161e+00 5.597e-01 1.000e+00
DQA1_199 A, *, T 3 2.025e+00 5.672e-01 1.000e+00
DPB1_194 *, R 2 1.111e+00 5.737e-01 1.000e+00
DPB1_215 *, I 2 1.111e+00 5.737e-01 1.000e+00
B_131 R, S 2 1.054e+00 5.902e-01 1.000e+00
A_107 G, W 2 1.047e+00 5.926e-01 1.000e+00
C_94 T, I 2 1.044e+00 5.935e-01 1.000e+00
A_63 N, E, Q 3 1.882e+00 5.973e-01 1.000e+00
DQA1_-6 M, * 2 1.031e+00 5.973e-01 1.000e+00
DRB1_26 F, Y, L 3 1.844e+00 6.054e-01 1.000e+00
A_74 D, H 2 1.002e+00 6.059e-01 1.000e+00
DQA1_187 A, *, T 3 1.819e+00 6.109e-01 1.000e+00
DQA1_215 F, *, L 3 1.819e+00 6.109e-01 1.000e+00
C_90 A, D 2 9.661e-01 6.169e-01 1.000e+00
C_95 L, I 2 9.617e-01 6.183e-01 1.000e+00
C_11 A, S 2 9.525e-01 6.211e-01 1.000e+00
B_158 A, T 2 9.012e-01 6.372e-01 1.000e+00
B_114 D, N, H 3 1.682e+00 6.410e-01 1.000e+00
A_-18 *, R 2 8.755e-01 6.455e-01 1.000e+00
A_-5 *, T 2 8.755e-01 6.455e-01 1.000e+00
A_284 *, I 2 8.755e-01 6.455e-01 1.000e+00
C_156 W, R, L, Q 4 2.473e+00 6.495e-01 1.000e+00
A_66 N, K 2 8.571e-01 6.515e-01 1.000e+00
A_142 I, T 2 8.559e-01 6.519e-01 1.000e+00
A_145 R, H 2 8.559e-01 6.519e-01 1.000e+00
A_-15 * 1 2.024e-01 6.528e-01 1.000e+00
A_294 * 1 2.024e-01 6.528e-01 1.000e+00
A_79 R, G 2 8.418e-01 6.565e-01 1.000e+00
A_80 I, T 2 8.418e-01 6.565e-01 1.000e+00
A_81 A, L 2 8.418e-01 6.565e-01 1.000e+00
A_82 L, R 2 8.418e-01 6.565e-01 1.000e+00
A_83 R, G 2 8.418e-01 6.565e-01 1.000e+00
DPB1_96 *, K, R 3 1.606e+00 6.581e-01 1.000e+00
DPB1_170 *, I, T 3 1.606e+00 6.581e-01 1.000e+00
A_99 Y, F 2 8.308e-01 6.601e-01 1.000e+00
B_177 D, E 2 8.213e-01 6.632e-01 1.000e+00
B_180 E, Q 2 8.213e-01 6.632e-01 1.000e+00
B_24 S, T, A 3 1.536e+00 6.740e-01 1.000e+00
DRB1_31 F, I 2 7.443e-01 6.893e-01 1.000e+00
B_156 R, L, W, D 4 2.250e+00 6.899e-01 1.000e+00
A_-11 *, S 2 7.305e-01 6.940e-01 1.000e+00
A_65 R, G 2 6.831e-01 7.107e-01 1.000e+00
A_-21 *, M 2 6.789e-01 7.122e-01 1.000e+00
DPB1_57 E, D 2 6.767e-01 7.129e-01 1.000e+00
A_95 I, V, L 3 1.360e+00 7.148e-01 1.000e+00
A_127 N, K 2 6.643e-01 7.174e-01 1.000e+00
C_177 E, K 2 6.392e-01 7.264e-01 1.000e+00
C_97 W, R 2 6.277e-01 7.306e-01 1.000e+00
A_44 R, K 2 6.239e-01 7.320e-01 1.000e+00
A_67 V, M 2 6.239e-01 7.320e-01 1.000e+00
A_150 A, V 2 6.239e-01 7.320e-01 1.000e+00
A_158 A, V 2 6.239e-01 7.320e-01 1.000e+00
A_276 *, P 2 6.091e-01 7.374e-01 1.000e+00
A_321 *, S 2 6.091e-01 7.374e-01 1.000e+00
DPB1_205 *, V, M 3 1.192e+00 7.550e-01 1.000e+00
A_298 *, I 2 5.482e-01 7.603e-01 1.000e+00
A_307 *, M 2 5.482e-01 7.603e-01 1.000e+00
DRB1_85 V, A 2 4.685e-01 7.912e-01 1.000e+00
A_246 *, A 2 4.583e-01 7.952e-01 1.000e+00
A_70 H, Q 2 4.549e-01 7.965e-01 1.000e+00
DQA1_25 Y, F 2 4.525e-01 7.975e-01 1.000e+00
B_95 L, W, I 3 9.848e-01 8.049e-01 1.000e+00
A_116 D, Y, H 3 8.461e-01 8.384e-01 1.000e+00
DPB1_65 I, L 2 3.524e-01 8.385e-01 1.000e+00
DPB1_35 F, L, Y 3 8.446e-01 8.388e-01 1.000e+00
C_6 R, K 2 3.499e-01 8.395e-01 1.000e+00
A_17 R, S 2 3.409e-01 8.433e-01 1.000e+00
A_62 R, G, E, Q, L 5 2.026e+00 8.456e-01 1.000e+00
C_9 Y, D, F, S 4 1.236e+00 8.722e-01 1.000e+00
A_299 *, T 2 2.288e-01 8.919e-01 1.000e+00
A_334 *, V 2 2.288e-01 8.919e-01 1.000e+00
DQB1_26 L, Y, G 3 6.006e-01 8.963e-01 1.000e+00
DQA1_34 Q, E 2 1.851e-01 9.116e-01 1.000e+00
A_152 E, V, A, R 4 9.529e-01 9.169e-01 1.000e+00
B_74 D, Y 2 1.542e-01 9.258e-01 1.000e+00
C_116 S, F, Y, L 4 7.424e-01 9.460e-01 1.000e+00
C_16 G, S 2 1.028e-01 9.499e-01 1.000e+00
A_97 R, M, I 3 2.072e-01 9.764e-01 1.000e+00
A_166 E, D 2 2.644e-02 9.869e-01 1.000e+00
A_167 W, G 2 2.644e-02 9.869e-01 1.000e+00
A_114 Q, H, R, E 4 2.952e-01 9.901e-01 1.000e+00
DPA1_91 * 0 0.000e+00 1.000e+00 1.000e+00
DPA1_111 * 0 0.000e+00 1.000e+00 1.000e+00
DPA1_127 * 0 0.000e+00 1.000e+00 1.000e+00
DPA1_160 * 0 0.000e+00 1.000e+00 1.000e+00
DPA1_190 * 0 0.000e+00 1.000e+00 1.000e+00
DPA1_228 * 0 0.000e+00 1.000e+00 1.000e+00
DQB1_-27 * 0 0.000e+00 1.000e+00 1.000e+00
DQB1_-21 * 0 0.000e+00 1.000e+00 1.000e+00
DQB1_-9 * 0 0.000e+00 1.000e+00 1.000e+00
DQB1_-6 * 0 0.000e+00 1.000e+00 1.000e+00
DQB1_-5 * 0 0.000e+00 1.000e+00 1.000e+00
DQB1_-4 * 0 0.000e+00 1.000e+00 1.000e+00
DQB1_203 * 0 0.000e+00 1.000e+00 1.000e+00
DQB1_220 * 0 0.000e+00 1.000e+00 1.000e+00
DQB1_221 * 0 0.000e+00 1.000e+00 1.000e+00
DQB1_224 * 0 0.000e+00 1.000e+00 1.000e+00
DRB1_-16 * 0 0.000e+00 1.000e+00 1.000e+00
DRB1_38 V, L 2 0.000e+00 1.000e+00 1.000e+00
DRB1_189 * 0 0.000e+00 1.000e+00 1.000e+00
DRB1_207 * 0 0.000e+00 1.000e+00 1.000e+00
DRB1_233 * 0 0.000e+00 1.000e+00 1.000e+00

Next, we can investigate how effect estimates are distributed for a given associated amino acid position, e.g. DQB1_9:

HLA_AA_DQB1_9_results <- runMiDAS(
  HLA_AA_model,
  experiment = "hla_aa",
  inheritance_model = "dominant",
  omnibus_groups_filter = "DQB1_9",
  lower_frequency_cutoff = 0.02,
  upper_frequency_cutoff = 0.98,
  correction = "bonferroni",
  exponentiate = TRUE
)

kableResults(HLA_AA_DQB1_9_results, scroll_box_height = "250px")
MiDAS analysis results
aa p.value p.adjusted estimate std.error conf.low conf.high statistic Ntotal Ntotal [%] N(disease=0) N(disease=0) [%] N(disease=1) N(disease=1) [%]
DQB1_9_F 7.000e-07 2.000e-06 2.0572 0.1451 1.5509 2.740 4.973 277 13.85% 103 10.30% 174 17.40%
DQB1_9_Y 1.103e-01 3.308e-01 0.5876 0.3329 0.2994 1.116 -1.597 960 48.00% 485 48.50% 475 47.50%
DQB1_9_* 1.000e+00 1.000e+00 1.0000 0.2167 0.6531 1.531 0.000 94 4.70% 47 4.70% 47 4.70%

This shows us that individuals carrying a Phenylalanine (F) at position 9 of DQB1 have a significantly increased risk, whereas individuals carrying a Tyrosine (Y) at the same amino acid position have a decreased risk.

It is logical to hypothesize that the risk residue is found on HLA-DQB1*06:02, the previously associated HLA risk allele. MiDAS thus provides the function getAllelesforAA to map amino acid residues back to the respective HLA alleles.

HLA_AA_DQB1_9_alleles <- getAllelesForAA(HLA_AA,"DQB1_9")
HLA-DQB1 (9) HLA-DQB1 alleles count frequency
05:03, 06:01 97 4.85%
F 04:01, 04:02, 04:23, 06:02 302 15.10%
Y 02:01, 02:02, 02:10, 03:01, 03:02, 03:03, 03:04, 03:05, 03:19, 03:22, 03:251, 03:96, 05:01, 05:02, 05:04, 05:107, 06:03, 06:04, 06:07, 06:09 1601 80.05%

Finally, it is also interesting to note that there are several amino acid positions coming up that determine the Bw4 binding motif (e.g. B_81), which is a determinant for interactions of HLA class I alleles with KIR on Natural Killer cells.

KIR associations and HLA-KIR interactions

Do we see association on the level of KIR genes, and when considering defined HLA-KIR interactions?

Now that we have performed KIR genotyping, or e.g. inferred KIR types from available whole-genome sequencing data, we can import this information, and check the gene frequencies. In our example, we could successfully infer KIR gene presence status for 935 out of the 1,000 individuals in our data set.

KIR data as imported by MiDAS
ID KIR3DL3 KIR2DS2 KIR2DL2 KIR2DL3 KIR2DP1 KIR2DL1 KIR3DP1 KIR2DL4 KIR3DL1 KIR3DS1 KIR2DL5 KIR2DS3 KIR2DS5 KIR2DS4 KIR2DS1 KIR3DL2
C001 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C002 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C003 1 0 0 1 1 1 1 1 0 1 1 0 1 0 1 1
C004 1 0 0 1 1 1 1 1 0 1 1 0 1 0 1 1
C005 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C006 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C007 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C008 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C009 1 1 1 0 1 1 1 1 0 1 1 1 1 0 1 1
C010 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1 1
C011 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C012 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
C013 1 1 1 1 1 0 1 1 0 1 1 0 1 0 1 1
C014 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
C015 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C016 1 0 0 1 1 1 1 1 0 1 1 0 1 0 1 1
C017 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1 1
C018 1 0 0 1 1 1 1 1 0 1 1 1 1 0 1 1
C019 1 0 0 1 1 1 1 1 0 1 1 0 1 0 1 1
C020 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1 1
C021 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
C022 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
C023 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
C024 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
C025 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
C026 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C027 1 1 1 0 1 1 1 1 0 1 1 1 1 0 1 1
C028 1 1 1 0 1 1 1 1 0 1 1 1 1 0 1 1
C029 1 0 0 1 1 1 1 1 0 1 1 1 1 0 1 1
C030 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1 1
C031 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
C032 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C033 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C034 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C035 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C036 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C037 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
C038 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C039 1 1 1 1 1 1 1 1 1 0 1 1 0 1 1 1
C040 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C041 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C042 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C043 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C044 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C045 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C046 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C047 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1 1
C048 1 1 1 0 1 1 1 1 0 1 1 1 1 0 1 1
C049 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1 1
C050 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
C051 1 1 1 1 1 0 1 1 0 1 1 0 1 0 1 1
C052 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
C053 1 0 0 1 1 1 1 1 0 1 1 0 1 0 1 1
C054 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1 1
C055 1 0 0 1 1 1 1 1 0 1 1 1 1 0 1 1
C056 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1 1
C057 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
C058 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
C059 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
C060 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
C061 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
C062 1 1 1 0 1 1 1 1 0 1 1 1 1 0 1 1
C063 1 1 1 0 1 1 1 1 0 1 1 1 1 0 1 1
C064 1 0 0 1 1 1 1 1 0 1 1 1 1 0 1 1
C065 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1 1
C066 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
C067 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C068 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C069 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C070 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
C071 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
C072 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1
C073 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C074 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C075 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C076 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
C077 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C078 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C079 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C080 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C081 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C082 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C083 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C084 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C085 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C086 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C087 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C088 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
C089 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1
C090 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
C091 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C092 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
C093 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C094 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
C095 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C096 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C097 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
C098 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C099 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C100 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C101 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1
C102 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C103 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C104 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C105 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C106 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C107 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1
C108 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C109 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
C110 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C111 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C112 1 1 1 1 1 0 1 1 1 0 0 0 0 1 1 1
C113 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C114 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
C115 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C116 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C117 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C118 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1 1
C119 1 1 1 0 1 1 1 1 1 1 1 0 1 1 1 1
C120 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
C121 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C122 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C123 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C124 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C125 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C126 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C127 1 1 1 0 0 0 1 1 1 0 0 0 0 1 0 1
C128 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C129 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C130 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C131 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C132 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C133 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C134 1 1 1 0 0 0 1 1 0 1 1 0 1 0 1 1
C135 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C136 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C137 1 0 0 1 1 1 1 1 1 0 0 0 0 1 1 1
C138 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C139 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
C140 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C141 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C142 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C143 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C144 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C145 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1
C146 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C147 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C148 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C149 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C150 1 1 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C151 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1
C152 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C153 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C154 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C155 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C156 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C157 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C158 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C159 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C160 1 0 0 1 1 1 1 1 0 1 1 0 1 0 1 1
C161 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C162 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C163 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C164 1 0 0 1 1 1 1 1 1 1 1 1 0 1 1 1
C165 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C166 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
C167 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C168 1 0 0 1 1 1 1 1 0 1 1 1 1 0 1 1
C169 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C170 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C171 1 0 0 1 1 1 1 1 1 1 1 1 0 1 1 1
C172 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
C173 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C174 1 0 0 1 1 1 1 1 1 1 1 1 0 1 1 1
C175 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C176 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C177 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C178 1 0 0 1 1 1 1 1 1 1 1 1 0 1 1 1
C179 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C180 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
C181 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C182 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C183 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C184 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C185 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C186 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C187 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C188 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C189 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C190 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
C191 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C192 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
C193 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C194 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C195 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C196 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C197 1 0 0 1 1 1 1 1 1 1 1 1 0 1 1 1
C198 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
C199 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
C200 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
C201 1 1 1 1 1 1 1 1 1 0 1 0 1 1 1 1
C202 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C203 1 0 0 1 1 1 1 1 1 0 0 0 0 1 1 1
C204 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C205 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
C206 1 1 1 1 1 1 1 1 1 1 1 1 0 1 0 1
C207 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C208 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C209 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C210 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C211 1 1 1 1 1 1 1 1 1 0 1 0 1 1 1 1
C212 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C213 1 1 1 1 1 1 1 1 0 1 1 1 0 0 1 1
C214 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C215 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
C216 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C217 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
C218 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C219 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C220 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C221 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C222 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C223 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C224 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C225 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C226 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C227 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C228 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
C229 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
C230 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C231 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C232 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C233 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C234 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C235 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
C236 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C237 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C238 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C239 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C240 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C241 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C242 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
C243 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C244 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C245 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C246 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C247 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C248 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C249 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C250 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C251 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C252 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C253 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C254 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C255 1 1 1 0 1 1 1 1 1 1 1 1 0 1 1 1
C256 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C257 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C258 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C259 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C260 1 0 0 1 1 1 1 1 1 0 0 0 0 1 1 1
C261 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C262 1 0 0 1 1 1 1 1 1 1 1 1 0 1 1 1
C263 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C264 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C265 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C266 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C267 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C268 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C269 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C270 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C271 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C272 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
C273 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C274 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C275 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C276 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C277 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C278 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C279 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C280 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C281 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C282 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C283 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C284 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1 1
C285 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C286 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C287 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C288 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C289 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
C290 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
C291 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C292 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C293 1 1 1 0 1 1 1 1 1 1 1 1 0 1 1 1
C294 1 1 1 0 0 0 1 1 1 0 0 0 0 1 0 1
C295 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C296 1 1 1 0 1 1 1 1 1 1 1 1 0 1 1 1
C297 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C298 1 1 1 1 1 0 1 1 1 0 0 0 0 1 1 1
C299 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C300 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C301 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C302 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
C303 1 1 1 0 1 1 1 1 1 1 1 1 0 1 1 1
C304 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C305 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C306 1 0 0 1 1 1 1 1 1 0 0 0 0 1 1 1
C307 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
C308 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
C309 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C310 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C311 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C312 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C313 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C314 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
C315 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
C316 1 0 0 1 1 1 1 1 1 1 1 1 0 1 1 1
C317 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C318 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C319 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C320 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C321 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
C322 1 0 0 1 1 1 1 1 0 1 1 0 1 0 1 1
C323 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C324 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C325 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C326 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C327 1 0 0 1 1 1 1 1 0 1 1 1 0 0 1 1
C328 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C329 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C330 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C331 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1
C332 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C333 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C334 1 0 0 1 1 1 1 1 1 1 1 1 0 1 1 1
C335 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C336 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C337 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C338 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C339 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
C340 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C341 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C342 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
C343 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C344 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C345 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C346 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C347 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C348 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C349 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C350 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C351 1 0 0 1 1 1 1 1 0 1 1 0 1 0 1 1
C352 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C353 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C354 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C355 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C356 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C357 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C358 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C359 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C360 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
C361 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
C362 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C363 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C364 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C365 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C366 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C367 1 0 0 1 1 1 1 1 1 0 0 0 0 1 1 1
C368 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
C369 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C370 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C371 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C372 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C373 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
C374 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C375 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C376 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
C377 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
C378 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C379 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C380 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C381 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C382 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C383 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
C384 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
C385 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C386 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C387 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C388 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C389 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C390 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C391 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C392 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
C393 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C394 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C395 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C396 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
C397 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
C398 1 0 0 1 1 1 1 1 0 1 1 1 1 0 1 1
C399 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C400 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C401 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C402 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C403 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
C404 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C405 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C406 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
C407 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C408 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C409 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C410 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C411 1 1 1 1 1 1 1 1 1 0 1 0 1 1 1 1
C412 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C413 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C414 1 0 0 1 1 1 1 1 0 1 1 0 1 0 1 1
C415 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C416 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C417 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
C418 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C419 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C420 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C421 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C422 1 0 0 1 1 1 1 1 0 1 1 0 1 0 1 1
C423 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
C424 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C425 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
C426 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C427 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C428 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C429 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C430 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C431 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C432 1 0 0 1 1 1 1 1 0 1 1 0 1 0 1 1
C433 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C434 1 1 1 0 1 1 1 1 1 1 1 1 0 1 1 1
C435 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C436 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C437 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C438 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1 1
C439 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C440 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C441 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
C442 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C443 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
C444 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C445 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
C446 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C447 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C448 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C449 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
C450 1 0 0 1 1 1 1 1 0 1 1 0 1 0 1 1
C451 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C452 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C453 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C454 1 1 1 1 1 1 1 1 1 0 1 0 1 1 1 1
C455 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C456 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C457 1 1 1 0 1 1 1 1 1 1 1 1 0 1 0 1
C458 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C459 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C460 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
C461 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C462 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C463 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
C464 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C465 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C466 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C467 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C468 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C469 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C470 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C471 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
C472 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C473 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C474 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C475 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C476 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C477 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C478 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C479 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C480 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
C481 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C482 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
C483 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C484 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C485 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C486 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C487 1 0 0 1 1 1 1 1 1 1 1 1 0 1 1 1
C488 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C489 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C490 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C491 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
C492 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C493 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C494 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
C495 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C496 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
C497 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
C498 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C499 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
C500 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P001 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P002 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P003 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P004 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P005 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P006 1 1 1 0 0 0 1 1 1 1 1 0 1 1 1 1
P007 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P008 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
P009 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P010 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P011 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P012 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P013 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P014 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P015 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P016 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P017 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P018 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P019 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P020 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P021 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P022 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P023 1 1 1 1 1 1 1 1 1 0 1 1 0 1 1 1
P024 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P025 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P026 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
P027 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P028 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
P029 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P030 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P031 1 1 1 0 0 0 1 1 1 0 1 0 1 1 1 1
P032 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
P033 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P034 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
P035 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P036 1 1 1 0 1 1 1 1 0 1 1 1 1 0 1 1
P037 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P038 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P039 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P040 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P041 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P042 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
P043 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P044 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P045 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P046 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P047 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P048 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
P049 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P050 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P051 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P052 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P053 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P054 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P055 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1 1
P056 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P057 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
P058 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P059 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
P060 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P061 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
P062 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P063 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P064 1 1 1 0 1 1 1 1 1 1 1 1 0 1 1 1
P065 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
P066 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P067 1 1 1 0 0 0 1 1 1 0 0 0 0 1 0 1
P068 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P069 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P070 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P071 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1
P072 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P073 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P074 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P075 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P076 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P077 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
P078 1 0 0 1 1 1 1 1 1 1 1 1 0 1 1 1
P079 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P080 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P081 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P082 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P083 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P084 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P085 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P086 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
P087 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P088 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P089 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P090 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P091 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P092 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P093 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P094 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P095 1 0 0 1 1 1 1 1 1 1 1 1 0 1 0 1
P096 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P097 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P098 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P099 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
P100 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P101 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P102 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P103 1 1 1 1 1 0 1 1 0 1 1 0 1 0 1 1
P104 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P105 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P106 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P107 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
P108 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P109 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
P110 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P111 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P112 1 1 1 1 1 0 1 1 1 0 0 0 0 1 1 1
P113 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P114 1 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1
P115 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P116 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P117 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P118 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P119 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P120 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P121 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
P122 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P123 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P124 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P125 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P126 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P127 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P128 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P129 1 1 1 0 0 0 1 1 1 1 1 0 1 1 1 1
P130 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P131 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P132 1 0 0 1 1 1 1 1 1 1 1 1 0 1 1 1
P133 1 0 0 1 1 1 1 1 0 1 1 0 1 0 1 1
P134 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P135 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P136 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P137 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P138 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1 1
P139 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P140 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P141 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P142 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P143 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
P144 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P145 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P146 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P147 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P148 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P149 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P150 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P151 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P152 1 1 1 0 1 1 1 1 1 1 1 1 0 1 1 1
P153 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P154 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P155 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P156 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P157 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P158 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P159 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P160 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P161 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P162 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P163 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P164 1 1 1 0 1 1 1 1 1 1 1 1 0 1 1 1
P165 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P166 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
P167 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P168 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P169 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
P170 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P171 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P172 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P173 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P174 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P175 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
P176 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P177 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P178 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P179 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P180 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P181 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
P182 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P183 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P184 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P185 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P186 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P187 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P188 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P189 1 0 0 1 1 1 1 1 0 1 1 1 1 0 1 1
P190 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P191 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P192 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P193 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1
P194 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P195 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P196 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
P197 1 1 1 0 0 0 1 1 1 1 1 0 1 1 1 1
P198 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P199 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
P200 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P201 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P202 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
P203 1 1 1 0 1 1 1 1 1 1 1 1 0 1 0 1
P204 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
P205 1 1 1 0 1 1 1 1 1 1 1 1 0 1 0 1
P206 1 1 1 0 1 1 1 1 1 1 1 1 0 1 1 1
P207 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P208 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P209 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P210 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P211 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
P212 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P213 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P214 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P215 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P216 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
P217 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P218 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P219 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P220 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P221 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P222 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P223 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P224 1 1 1 1 1 1 1 1 1 1 1 1 0 1 0 1
P225 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
P226 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P227 1 1 1 1 1 1 1 1 1 0 1 0 1 1 1 1
P228 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P229 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P230 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P231 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1 1
P232 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P233 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P234 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P235 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P236 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P237 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P238 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P239 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P240 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
P241 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P242 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P243 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P244 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P245 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P246 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
P247 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P248 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P249 1 1 1 0 1 1 1 1 1 1 1 1 0 1 1 1
P250 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P251 1 0 0 1 1 1 1 1 1 1 1 1 0 1 0 1
P252 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P253 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
P254 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P255 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P256 1 1 1 0 1 1 1 1 1 1 1 1 0 1 0 1
P257 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P258 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
P259 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
P260 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P261 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P262 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
P263 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P264 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1
P265 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P266 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
P267 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
P268 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
P269 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
P270 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P271 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
P272 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
P273 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P274 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P275 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
P276 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P277 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
P278 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
P279 1 1 1 1 1 1 1 1 1 1 1 1 0 1 0 1
P280 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P281 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P282 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P283 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
P284 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P285 1 1 1 1 1 0 1 1 1 0 0 0 0 1 1 1
P286 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P287 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P288 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P289 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
P290 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P291 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P292 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P293 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P294 1 0 0 1 1 1 1 1 1 1 1 1 0 1 1 1
P295 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P296 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P297 1 1 1 0 1 1 1 1 1 1 1 0 1 1 1 1
P298 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P299 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P300 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P301 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
P302 1 1 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P303 1 1 1 0 1 1 1 1 1 1 1 1 0 1 1 1
P304 1 0 0 1 1 1 1 1 1 1 1 1 0 1 0 1
P305 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
P306 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P307 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P308 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P309 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
P310 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P311 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P312 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P313 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P314 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P315 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P316 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P317 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P318 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
P319 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P320 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P321 1 0 0 1 1 1 1 1 1 1 1 1 0 1 1 1
P322 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P323 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
P324 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
P325 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P326 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P327 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
P328 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P329 1 1 1 0 1 1 1 1 0 1 1 1 1 0 1 1
P330 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P331 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P332 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P333 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P334 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P335 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P336 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P337 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P338 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1
P339 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1
P340 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P341 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P342 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P343 1 1 1 0 0 0 1 1 1 1 1 0 1 1 1 1
P344 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P345 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P346 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P347 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
P348 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P349 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P350 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
P351 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P352 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
P353 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P354 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P355 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P356 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P357 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P358 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P359 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P360 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P361 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
P362 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P363 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P364 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P365 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
P366 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P367 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P368 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P369 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P370 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P371 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P372 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P373 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P374 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P375 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P376 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P377 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P378 1 1 1 0 1 1 1 1 1 1 1 1 0 1 1 1
P379 1 1 1 0 1 1 1 1 0 1 1 1 1 0 1 1
P380 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P381 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P382 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P383 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
P384 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P385 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P386 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P387 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P388 1 1 1 0 1 1 1 1 1 1 1 1 0 1 1 1
P389 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P390 1 0 0 1 1 1 1 1 1 0 0 0 0 1 1 1
P391 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P392 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
P393 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P394 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P395 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P396 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P397 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P398 1 0 0 1 1 1 1 1 1 1 1 1 0 1 1 1
P399 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P400 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P401 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P402 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P403 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P404 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P405 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P406 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P407 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P408 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P409 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
P410 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
P411 1 1 1 0 1 1 1 1 1 1 1 1 0 1 1 1
P412 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P413 1 0 0 1 1 1 1 1 1 1 1 1 0 1 1 1
P414 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P415 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P416 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P417 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P418 1 1 1 1 1 1 1 1 1 1 1 1 0 1 0 1
P419 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P420 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P421 1 0 0 1 1 1 1 1 0 1 1 1 1 0 1 1
P422 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P423 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P424 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P425 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1
P426 1 0 0 1 1 1 1 1 1 1 1 1 0 1 1 1
P427 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P428 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P429 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P430 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P431 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
P432 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P433 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P434 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P435 1 1 1 1 1 1 1 1 1 1 1 1 0 1 0 1
P436 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P437 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P438 1 0 0 1 1 1 1 1 1 0 0 0 0 1 1 1
P439 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
P440 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P441 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P442 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1 1
P443 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P444 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P445 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P446 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P447 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P448 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P449 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P450 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P451 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P452 1 1 1 1 1 1 1 1 1 1 1 1 0 1 0 1
P453 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P454 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P455 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
P456 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P457 1 1 1 1 1 1 1 1 1 1 1 1 0 1 0 1
P458 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P459 1 1 1 0 1 1 1 1 1 1 1 1 0 1 1 1
P460 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P461 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P462 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1
P463 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P464 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P465 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P466 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P467 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1
P468 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P469 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P470 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1
P471 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
P472 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P473 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P474 1 1 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P475 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
P476 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P477 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P478 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P479 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P480 1 1 1 0 0 0 1 1 1 0 0 0 0 1 0 1
P481 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P482 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA
P483 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P484 1 1 1 0 1 1 1 1 1 0 1 1 0 1 0 1
P485 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1
P486 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P487 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P488 1 1 1 0 1 1 1 1 1 1 1 1 0 1 1 1
P489 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P490 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P491 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P492 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P493 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1
P494 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1
P495 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P496 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P497 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1
P498 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
P499 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1
P500 1 1 1 1 1 1 1 1 1 0 1 0 1 1 1 1

Next, we want to check our KIR genes frequencies.

KIR data as imported by MiDAS
gene Counts Freq
KIR3DL3 935 1.0000000
KIR2DS2 455 0.4866310
KIR2DL2 449 0.4802139
KIR2DL3 849 0.9080214
KIR2DP1 925 0.9893048
KIR2DL1 918 0.9818182
KIR3DP1 935 1.0000000
KIR2DL4 935 1.0000000
KIR3DL1 853 0.9122995
KIR3DS1 365 0.3903743
KIR2DL5 485 0.5187166
KIR2DS3 294 0.3144385
KIR2DS5 288 0.3080214
KIR2DS4 853 0.9122995
KIR2DS1 371 0.3967914
KIR3DL2 935 1.0000000

Next, we rerun our prepareMiDAS function, this time including the KIR data. We prepare the data to test for both KIR gene associations, as well as HLA-KIR interactions. But first, let’s runMiDAS on the level of KIR genes.

MiDAS analysis results
kir.gene p.value p.adjusted estimate std.error conf.low conf.high statistic Ntotal Ntotal [%] N(disease=0) N(disease=0) [%] N(disease=1) N(disease=1) [%]
KIR3DL1 3.700e-06 3.650e-05 3.4291 0.2661 2.0717 5.9126 4.6303 853 91.23% 405 86.72% 448 95.73%
KIR2DS4 3.700e-06 3.650e-05 3.4291 0.2661 2.0717 5.9126 4.6303 853 91.23% 405 86.72% 448 95.73%
KIR2DL3 4.414e-02 4.414e-01 0.6282 0.2310 0.3965 0.9835 -2.0128 849 90.80% 433 92.72% 416 88.89%
KIR2DS2 8.295e-02 8.295e-01 1.2552 0.1311 0.9710 1.6235 1.7338 455 48.66% 214 45.82% 241 51.50%
KIR2DL2 1.087e-01 1.000e+00 1.2341 0.1311 0.9546 1.5962 1.6043 449 48.02% 212 45.40% 237 50.64%
KIR2DS5 1.143e-01 1.000e+00 0.7992 0.1420 0.6046 1.0552 -1.5790 288 30.80% 155 33.19% 133 28.42%
KIR2DS3 1.658e-01 1.000e+00 1.2160 0.1411 0.9224 1.6044 1.3858 294 31.44% 137 29.34% 157 33.55%
KIR2DS1 1.949e-01 1.000e+00 0.8407 0.1338 0.6465 1.0927 -1.2962 371 39.68% 195 41.76% 176 37.61%
KIR3DS1 7.178e-01 1.000e+00 0.9527 0.1341 0.7324 1.2391 -0.3614 365 39.04% 185 39.61% 180 38.46%
KIR2DL5 9.208e-01 1.000e+00 0.9871 0.1309 0.7636 1.2759 -0.0994 485 51.87% 243 52.03% 242 51.71%

We found an association of KIR3DL1 gene presence with disease status. Since HLA alleles encoding Bw4 epitopes are ligands for KIR3DL1, we can now ask the question whether patients coding for both receptor and ligand are at increased risk for our disease phenotype.

HLA-KIR interactions

Do known biological interactions between KIR receptors and their HLA ligands show significant assocation?

If both HLA alleles and KIR gene presence / absence information is available, MiDAS can encode experimentally validated receptor-ligand interactions, defined according to Pende et al. Please note that there is extensive allelic variation in KIR genes, as well as evidence that their interaction with HLA does not only depend on the presence of the KIR gene, but also allele status. Currently, MiDAS does not consider allelic variation in KIR, but a custom dictionary can be provided by the user (see documentation).

First let’s explore HLA-KIR interactions frequencies

HLA-KIR interaction frequencies
term Counts Freq
C2_KIR2DL1 580 0.6079665
C1_KIR2DL2 373 0.3922187
B*46:01_KIR2DL2 4 0.0040000
B*73:01_KIR2DL2 0 0.0000000
C2_KIR2DL2 283 0.2966457
C1_KIR2DL3 704 0.7402734
B*46:01_KIR2DL3 10 0.0100000
B*73:01_KIR2DL3 2 0.0020000
C2_KIR2DL3 537 0.5628931
A*23_KIR3DL1 36 0.0363269
A*24_KIR3DL1 196 0.1983806
A*32_KIR3DL1 54 0.0541625
A*03_KIR3DL2 189 0.1926606
A*11_KIR3DL2 116 0.1161161
Bw4_KIR3DL1 680 0.7188161
Bw4(HLA-B_only)_KIR3DL1 585 0.6138510
C2_KIR2DS1 237 0.2484277
C1_KIR2DS2 379 0.3985279
A*11:01_KIR2DS2 48 0.0480480
C*02:02_KIR2DS4 76 0.0763052
C*04:01_KIR2DS4 180 0.1832994
C*05:01_KIR2DS4 122 0.1238579
C*01:02_KIR2DS4 85 0.0855131
C*14:02_KIR2DS4 28 0.0280280
C*16:01_KIR2DS4 42 0.0421264
A*11_KIR2DS4 102 0.1021021
C2_KIR2DS5 178 0.1865828
B*51_KIR3DS1 53 0.0532663
A03_A11_KIR2DS2 126 0.1285714
MiDAS analysis results
hla.kir.interaction p.value p.adjusted estimate std.error conf.low conf.high statistic Ntotal Ntotal [%] N(disease=0) N(disease=0) [%] N(disease=1) N(disease=1) [%]
Bw4_KIR3DL1 6.000e-07 1.620e-05 2.0952 0.1487 1.5686 2.810 4.976e+00 680 71.88% 306 64.56% 374 79.24%
Bw4(HLA-B_only)_KIR3DL1 6.240e-05 1.559e-03 1.7131 0.1345 1.3173 2.232 4.004e+00 585 61.39% 262 55.04% 323 67.71%
C1_KIR2DS2 1.895e-03 4.737e-02 1.5127 0.1332 1.1656 1.966 3.106e+00 379 39.85% 167 34.94% 212 44.82%
C1_KIR2DL2 2.877e-03 7.193e-02 1.4889 0.1336 1.1466 1.936 2.981e+00 373 39.22% 165 34.52% 208 43.97%
B*51_KIR3DS1 1.801e-02 4.502e-01 2.0202 0.2973 1.1427 3.692 2.365e+00 53 5.33% 18 3.61% 35 7.04%
A*24_KIR3DL1 1.111e-01 1.000e+00 1.2909 0.1602 0.9437 1.770 1.593e+00 196 19.84% 88 17.81% 108 21.86%
C2_KIR2DS5 1.352e-01 1.000e+00 0.7793 0.1669 0.5610 1.080 -1.494e+00 178 18.66% 98 20.55% 80 16.77%
C2_KIR2DS1 1.549e-01 1.000e+00 0.8076 0.1502 0.6011 1.084 -1.422e+00 237 24.84% 128 26.83% 109 22.85%
A*11_KIR3DL2 1.712e-01 1.000e+00 0.7618 0.1988 0.5143 1.123 -1.368e+00 116 11.61% 65 13.00% 51 10.22%
A*11:01_KIR2DS2 2.415e-01 1.000e+00 0.7038 0.2999 0.3862 1.260 -1.171e+00 48 4.80% 28 5.60% 20 4.01%
A*03_KIR3DL2 3.149e-01 1.000e+00 1.1771 0.1622 0.8568 1.619 1.005e+00 189 19.27% 88 18.00% 101 20.53%
C*02:02_KIR2DS4 3.498e-01 1.000e+00 1.2515 0.2399 0.7832 2.013 9.351e-01 76 7.63% 34 6.84% 42 8.42%
A*03_A*11_KIR2DS2 3.529e-01 1.000e+00 1.1946 0.1914 0.8215 1.742 9.290e-01 126 12.86% 58 11.86% 68 13.85%
A*11_KIR2DS4 4.096e-01 1.000e+00 0.8413 0.2096 0.5563 1.268 -8.245e-01 102 10.21% 55 11.00% 47 9.42%
C*05:01_KIR2DS4 4.465e-01 1.000e+00 1.1590 0.1938 0.7931 1.698 7.613e-01 122 12.39% 57 11.59% 65 13.18%
C2_KIR2DL3 4.728e-01 1.000e+00 0.9105 0.1306 0.7048 1.176 -7.179e-01 537 56.29% 274 57.44% 263 55.14%
A*32_KIR3DL1 5.709e-01 1.000e+00 0.8530 0.2806 0.4890 1.478 -5.668e-01 54 5.42% 29 5.82% 25 5.01%
C*04:01_KIR2DS4 5.996e-01 1.000e+00 1.0905 0.1650 0.7891 1.508 5.250e-01 180 18.33% 87 17.68% 93 18.98%
C2_KIR2DL2 6.198e-01 1.000e+00 1.0729 0.1418 0.8126 1.417 4.961e-01 283 29.66% 138 28.93% 145 30.40%
C*01:02_KIR2DS4 7.047e-01 1.000e+00 1.0898 0.2270 0.6982 1.705 3.790e-01 85 8.55% 41 8.22% 44 8.89%
C*14:02_KIR2DS4 7.057e-01 1.000e+00 0.8649 0.3843 0.4012 1.840 -3.777e-01 28 2.80% 15 3.00% 13 2.61%
A*23_KIR3DL1 7.483e-01 1.000e+00 1.1153 0.3401 0.5717 2.193 3.209e-01 36 3.63% 17 3.44% 19 3.82%
C*16:01_KIR2DS4 7.576e-01 1.000e+00 1.1023 0.3157 0.5928 2.061 3.086e-01 42 4.21% 20 4.02% 22 4.41%
C2_KIR2DL1 8.945e-01 1.000e+00 0.9826 0.1326 0.7575 1.274 -1.326e-01 580 60.80% 291 61.01% 289 60.59%
C1_KIR2DL3 9.824e-01 1.000e+00 0.9967 0.1479 0.7458 1.332 -2.209e-02 704 74.03% 354 74.06% 350 74.00%

Patients who are carriers of both HLA-Bw4 and KIR3DL1 are at a significantly increased risk for having the disease, which is consistent with the associations we saw on KIR ligand and KIR gene level.

Finally, let’s now take a look at some additional ways to analyze immunogenetics data within MiDAS, which are probably not relevant for most users.

HLA heterozygosity and evolutionary divergence

Instead of focusing on single HLA alleles or their interactions, you might be interested in HLA diversity. For example, it has been shown that MHC heterozygosity confers a selective advantage against infections with multiple Salmonella strains. But given the degree of variability in HLA genes, heterozygosity is a rather crude measure. On the amino acid sequence level, it is possible to perform more refined analyses, by calculating the level of evolutionary divergence between pairs of HLA alleles of different genes. Pierini and Lenz have found Grantham’s distance score to be a good proxy for HLA functional divergence. It has been demonstrated that HLA-C allelic divergence was associated with HIV viral load. In addition, there is evidence for a role of HLA class I divergence in the efficacy of cancer immunotherapy.

MiDAS provides experiment types hla_het to analyze heterozygosity of class I and II genes, as well as hla_divergence to analyze classical class I genes using Grantham’s distance:

HLA_het <- prepareMiDAS(
  hla_calls = dat_HLA, 
  colData = dat_pheno, 
  experiment = c("hla_het","hla_divergence")
)

HLA_het_model <- glm(outcome ~ term, data=HLA_het, family=binomial())

HLA_het_results <- runMiDAS(HLA_het_model, 
  experiment = "hla_het", 
  exponentiate = TRUE
)

kableResults(HLA_het_results)
MiDAS analysis results
term p.value p.adjusted estimate std.error conf.low conf.high statistic Ntotal Ntotal [%] N(outcome=0) N(outcome=0) [%] N(outcome=1) N(outcome=1) [%]
C_het 5.405e-02 0.4865 0.5436 0.3165 0.2878 1.003 -1.9264 900 90.00% 246 93.18% 208 88.14%
DPA1_het 2.155e-01 1.0000 1.2585 0.1856 0.8748 1.812 1.2386 350 35.00% 91 34.47% 94 39.83%
DQA1_het 4.297e-01 1.0000 1.2477 0.2802 0.7230 2.179 0.7897 868 86.80% 230 87.12% 211 89.41%
B_het 4.672e-01 1.0000 1.2998 0.3606 0.6455 2.686 0.7271 934 93.40% 244 92.42% 222 94.07%
A_het 5.067e-01 1.0000 1.2057 0.2817 0.6962 2.111 0.6640 892 89.20% 231 87.50% 211 89.41%
DPB1_het 5.279e-01 1.0000 1.1468 0.2170 0.7504 1.759 0.6312 779 77.90% 203 76.89% 187 79.24%
DRB1_het 6.693e-01 1.0000 0.8824 0.2931 0.4954 1.572 -0.4271 890 89.00% 238 90.15% 210 88.98%
DRA_het 6.997e-01 1.0000 0.9329 0.1800 0.6553 1.327 -0.3857 435 43.50% 122 46.21% 105 44.49%
DQB1_het 7.659e-01 1.0000 0.9184 0.2860 0.5234 1.614 -0.2977 882 88.20% 236 89.39% 209 88.56%
HLA_div_results <- runMiDAS(HLA_het_model, 
  experiment = "hla_divergence", 
  exponentiate = TRUE
)

kableResults(HLA_div_results, scroll_box_height = "250px")
MiDAS analysis results
term p.value p.adjusted estimate std.error conf.low conf.high statistic
ABC_avg 0.3713 1 1.043 4.668e-02 0.9517 1.143 0.8940
A 0.4519 1 1.018 2.416e-02 0.9713 1.068 0.7522
B 0.6163 1 1.013 2.577e-02 0.9632 1.066 0.5012
C 0.7747 1 1.012 4.052e-02 0.9345 1.096 0.2863

In our example dataset, our results show that there is no significant association of HLA heterozygosity or evolutionary divergence with our disease phenotype of interest.

We hope that this tutorial is useful to get you started working with MiDAS. Please don’t hesitate to contact us if you have questions or suggestions for improvement.