Breast survival dataset using network from STRING DB
2024-02-05
Contents
1 Instalation
if (!require("BiocManager"))
install.packages("BiocManager")
BiocManager::install("glmSparseNet")2 Required Packages
library(dplyr)
library(Matrix)
library(ggplot2)
library(forcats)
library(parallel)
library(reshape2)
library(survival)
library(VennDiagram)
library(futile.logger)
library(curatedTCGAData)
library(TCGAutils)
library(glmSparseNet)
#
# Some general options for futile.logger the debugging package
.Last.value <- flog.layout(layout.format('[~l] ~m'))
.Last.value <- glmSparseNet:::show.message(FALSE)
# Setting ggplot2 default theme as minimal
theme_set(ggplot2::theme_minimal())3 Overview
This vignette uses the STRING database (https://string-db.org/) of protein-protein interactions as the network-based penalizer in generalized linear models using Breast invasive carcinoma sample dataset.
The degree vector is calculated manually to account for genes that are not present in the STRING database, as these will not have any interactions, i.e. edges.
3.1 Download Data from STRING
Retrieve all interactions from STRING databse. We have included a helper function that retrieves the Homo sapiens known interactions.
For this vignette, we use a cached version of all interaction with
score_threshold = 700
Note: Text-based interactions are excluded from the network.
# Not evaluated in vignette as it takes too long to download and process
all.interactions.700 <- stringDBhomoSapiens(score_threshold = 700)
string.network <- buildStringNetwork(all.interactions.700,
use.names = 'external')4 Build network matrix
Build a sparse matrix object that contains the network.
string.network.undirected <- string.network + Matrix::t(string.network)
string.network.undirected <- (string.network.undirected != 0) * 15 Network Statistics
5.1 Graph information
## [INFO] Directed graph (score_threshold = 700)
## [INFO] * total edges: 225330
## [INFO] * unique protein: 11033
## [INFO] * edges per protein: 20.423276
## [INFO]
## [INFO] Undirected graph (score_threshold = 700)
## [INFO] * total edges: 225209
## [INFO] * unique protein: 11033
## [INFO] * edges per protein: 20.4123095.2 Summary of degree (indegree + outdegree)
## [INFO] Summary of degree:
##
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1.00 3.00 13.00 40.83 43.00 4175.005.3 Histogram of degree (up until 99.999% quantile)
qplot(degree.network[degree.network <= quantile(degree.network,
probs = .99999)],
geom = 'histogram', fill = my.colors(2), bins = 100) +
theme(legend.position = 'none') + xlab('Degree (up until 99.999% quantile)')
6 glmSparseNet
- Dataset from curatedTCGAdata
brca <- curatedTCGAData(diseaseCode = "BRCA", assays = "RNASeq2GeneNorm",
version = "1.1.38", dry.run = FALSE)Build the survival data from the clinical columns.
- Selects only primary solid tumour samples
- Merge survival times for patients, which have different columns in case they are alive or dead.
- Build two matrix objects that fit the data
xdataandydata
# keep only solid tumour (code: 01)
brca.primary.solid.tumor <- TCGAutils::TCGAsplitAssays(brca, '01')
xdata.raw <- t(assay(brca.primary.solid.tumor[[1]]))
# Get survival information
ydata.raw <- colData(brca.primary.solid.tumor) %>% as.data.frame %>%
# Convert days to integer
dplyr::mutate(
Days.to.date.of.Death = as.integer(Days.to.date.of.Death),
Days.to.Last.Contact = as.integer(Days.to.Date.of.Last.Contact)
) %>%
# Find max time between all days (ignoring missings)
dplyr::rowwise() %>%
dplyr::mutate(
time = max(days_to_last_followup, Days.to.date.of.Death,
Days.to.Last.Contact, days_to_death, na.rm = TRUE)
) %>%
# Keep only survival variables and codes
dplyr::select(patientID, status = vital_status, time) %>%
# Discard individuals with survival time less or equal to 0
dplyr::filter(!is.na(time) & time > 0) %>%
as.data.frame()
# Set index as the patientID
rownames(ydata.raw) <- ydata.raw$patientID
# keep only features that are in degree.network and have standard deviation > 0
valid.features <- colnames(xdata.raw)[colnames(xdata.raw) %in%
names(degree.network[degree.network > 0])]
xdata.raw <- xdata.raw[TCGAbarcode(rownames(xdata.raw)) %in%
rownames(ydata.raw), valid.features]
xdata.raw <- scale(xdata.raw)
# Order ydata the same as assay
ydata.raw <- ydata.raw[TCGAbarcode(rownames(xdata.raw)), ]
# Using only a subset of genes previously selected to keep this short example.
set.seed(params$seed)
small.subset <- c('AAK1', 'ADRB1', 'AK7', 'ALK', 'APOBEC3F', 'ARID1B', 'BAMBI',
'BRAF', 'BTG1', 'CACNG8', 'CASP12', 'CD5', 'CDA', 'CEP72',
'CPD', 'CSF2RB', 'CSN3', 'DCT', 'DLG3', 'DLL3', 'DPP4',
'DSG1', 'EDA2R', 'ERP27', 'EXD1', 'GABBR2', 'GADD45A',
'GBP1', 'HTR1F', 'IFNK', 'IRF2', 'IYD', 'KCNJ11', 'KRTAP5-6',
'MAFA', 'MAGEB4', 'MAP2K6', 'MCTS1', 'MMP15', 'MMP9',
'NFKBIA', 'NLRC4', 'NT5C1A', 'OPN4', 'OR13C5', 'OR13C8',
'OR2T6', 'OR4K2', 'OR52E6', 'OR5D14', 'OR5H1', 'OR6C4',
'OR7A17', 'OR8J3', 'OSBPL1A', 'PAK6', 'PDE11A', 'PELO',
'PGK1', 'PIK3CB', 'PMAIP1', 'POLR2B', 'POP1', 'PPFIA3',
'PSME1', 'PSME2', 'PTEN', 'PTGES3', 'QARS', 'RABGAP1',
'RBM3', 'RFC3', 'RGPD8', 'RPGRIP1L', 'SAV1', 'SDC1', 'SDC3',
'SEC16B', 'SFPQ', 'SFRP5', 'SIPA1L1', 'SLC2A14', 'SLC6A9',
'SPATA5L1', 'SPINT1', 'STAR', 'STXBP5', 'SUN3', 'TACC2',
'TACR1', 'TAGLN2', 'THPO', 'TNIP1', 'TP53', 'TRMT2B', 'TUBB1',
'VDAC1', 'VSIG8', 'WNT3A', 'WWOX', 'XRCC4', 'YME1L1',
'ZBTB11', 'ZSCAN21') %>%
sample(size = 50) %>% sort
# make sure we have 100 genes
small.subset <- c(small.subset, sample(colnames(xdata.raw), 51)) %>%
unique %>%
sort
xdata <- xdata.raw[, small.subset[small.subset %in% colnames(xdata.raw)]]
ydata <- ydata.raw %>%
dplyr::select(time, status) %>%
dplyr::filter(!is.na(time) | time < 0)- Build dataset that overlaps with STRING data
#
# Add degree 0 to genes not in STRING network
my.degree <- degree.network[small.subset]
my.string <- string.network.binary[small.subset, small.subset]Degree distribution for sample set of gene features (in xdata).
6.1 Select balanced folds for cross-validation
set.seed(params$seed)
foldid <- glmSparseNet:::balanced.cv.folds(!!ydata$status)$output6.2 glmHub model
Penalizes using the Hub heuristics, see hubHeuristic function definition for
more details.
cv.hub <- cv.glmHub(xdata,
Surv(ydata$time, ydata$status),
family = 'cox',
foldid = foldid,
network = my.string,
network.options = networkOptions(min.degree = 0.2))Kaplan-Meier estimator separating individuals by low and high risk (based on model’s coefficients)
## $pvalue
## [1] 9.628346e-05
##
## $plot
##
## $km
## Call: survfit(formula = survival::Surv(time, status) ~ group, data = prognostic.index.df)
##
## n events median 0.95LCL 0.95UCL
## Low risk 983 129 3959 3738 7455
## High risk 97 23 2911 2469 NA6.3 glmOrphan model
Penalizes using the Orphan heuristics, see orphanHeuristic function
definition for more details.
cv.orphan <- cv.glmOrphan(xdata,
Surv(ydata$time, ydata$status),
family = 'cox',
foldid = foldid,
network = my.string,
network.options = networkOptions(min.degree = 0.2))Kaplan-Meier estimator separating individuals by low and high risk (based on model’s coefficients)
## $pvalue
## [1] 0
##
## $plot
##
## $km
## Call: survfit(formula = survival::Surv(time, status) ~ group, data = prognostic.index.df)
##
## n events median 0.95LCL 0.95UCL
## Low risk 540 34 7455 6456 NA
## High risk 540 118 2866 2573 34726.4 Elastic Net model (without network-penalization)
Uses regular glmnet model as simple baseline
cv.glm <- cv.glmnet(xdata,
Surv(ydata$time, ydata$status),
family = 'cox',
foldid = foldid)Kaplan-Meier estimator separating individuals by low and high risk (based on model’s coefficients)
## $pvalue
## [1] 0
##
## $plot
##
## $km
## Call: survfit(formula = survival::Surv(time, status) ~ group, data = prognostic.index.df)
##
## n events median 0.95LCL 0.95UCL
## Low risk 540 38 6593 4398 NA
## High risk 540 114 2854 2551 34616.5 Selected genes
Venn diagram of overlapping genes.
Descriptive table showing which genes are selected in each model
We can observe, that elastic net without network-based penalization selects the best model with 40% more genes than glmOrphan and glmHub, without loosing accuracy.
note: size of circles represent the degree of that gene in network.
7 Session Info
sessionInfo()## R version 4.3.2 Patched (2023-11-13 r85521)
## Platform: x86_64-pc-linux-gnu (64-bit)
## Running under: Ubuntu 22.04.3 LTS
##
## Matrix products: default
## BLAS: /home/biocbuild/bbs-3.18-bioc/R/lib/libRblas.so
## LAPACK: /usr/lib/x86_64-linux-gnu/lapack/liblapack.so.3.10.0
##
## locale:
## [1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C
## [3] LC_TIME=en_GB LC_COLLATE=C
## [5] LC_MONETARY=en_US.UTF-8 LC_MESSAGES=en_US.UTF-8
## [7] LC_PAPER=en_US.UTF-8 LC_NAME=C
## [9] LC_ADDRESS=C LC_TELEPHONE=C
## [11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C
##
## time zone: America/New_York
## tzcode source: system (glibc)
##
## attached base packages:
## [1] grid parallel stats4 stats graphics grDevices utils
## [8] datasets methods base
##
## other attached packages:
## [1] VennDiagram_1.7.3 reshape2_1.4.4
## [3] forcats_1.0.0 glmSparseNet_1.20.1
## [5] glmnet_4.1-8 Matrix_1.6-5
## [7] TCGAutils_1.22.2 curatedTCGAData_1.24.0
## [9] MultiAssayExperiment_1.28.0 SummarizedExperiment_1.32.0
## [11] Biobase_2.62.0 GenomicRanges_1.54.1
## [13] GenomeInfoDb_1.38.5 IRanges_2.36.0
## [15] S4Vectors_0.40.2 BiocGenerics_0.48.1
## [17] MatrixGenerics_1.14.0 matrixStats_1.2.0
## [19] futile.logger_1.4.3 survival_3.5-7
## [21] ggplot2_3.4.4 dplyr_1.1.4
## [23] BiocStyle_2.30.0
##
## loaded via a namespace (and not attached):
## [1] jsonlite_1.8.8 shape_1.4.6
## [3] magrittr_2.0.3 magick_2.8.2
## [5] GenomicFeatures_1.54.3 farver_2.1.1
## [7] rmarkdown_2.25 BiocIO_1.12.0
## [9] zlibbioc_1.48.0 vctrs_0.6.5
## [11] memoise_2.0.1 Rsamtools_2.18.0
## [13] RCurl_1.98-1.14 rstatix_0.7.2
## [15] htmltools_0.5.7 S4Arrays_1.2.0
## [17] BiocBaseUtils_1.4.0 progress_1.2.3
## [19] AnnotationHub_3.10.0 lambda.r_1.2.4
## [21] curl_5.2.0 broom_1.0.5
## [23] pROC_1.18.5 SparseArray_1.2.3
## [25] sass_0.4.8 bslib_0.6.1
## [27] plyr_1.8.9 zoo_1.8-12
## [29] futile.options_1.0.1 cachem_1.0.8
## [31] GenomicAlignments_1.38.2 mime_0.12
## [33] lifecycle_1.0.4 iterators_1.0.14
## [35] pkgconfig_2.0.3 R6_2.5.1
## [37] fastmap_1.1.1 GenomeInfoDbData_1.2.11
## [39] shiny_1.8.0 digest_0.6.34
## [41] colorspace_2.1-0 AnnotationDbi_1.64.1
## [43] ExperimentHub_2.10.0 RSQLite_2.3.5
## [45] ggpubr_0.6.0 filelock_1.0.3
## [47] labeling_0.4.3 km.ci_0.5-6
## [49] fansi_1.0.6 httr_1.4.7
## [51] abind_1.4-5 compiler_4.3.2
## [53] bit64_4.0.5 withr_3.0.0
## [55] backports_1.4.1 BiocParallel_1.36.0
## [57] carData_3.0-5 DBI_1.2.1
## [59] highr_0.10 ggsignif_0.6.4
## [61] biomaRt_2.58.2 rappdirs_0.3.3
## [63] DelayedArray_0.28.0 rjson_0.2.21
## [65] tools_4.3.2 interactiveDisplayBase_1.40.0
## [67] httpuv_1.6.14 glue_1.7.0
## [69] restfulr_0.0.15 promises_1.2.1
## [71] generics_0.1.3 gtable_0.3.4
## [73] KMsurv_0.1-5 tzdb_0.4.0
## [75] tidyr_1.3.1 survminer_0.4.9
## [77] data.table_1.15.0 hms_1.1.3
## [79] car_3.1-2 xml2_1.3.6
## [81] utf8_1.2.4 XVector_0.42.0
## [83] BiocVersion_3.18.1 foreach_1.5.2
## [85] pillar_1.9.0 stringr_1.5.1
## [87] later_1.3.2 splines_4.3.2
## [89] BiocFileCache_2.10.1 lattice_0.22-5
## [91] rtracklayer_1.62.0 bit_4.0.5
## [93] tidyselect_1.2.0 Biostrings_2.70.2
## [95] knitr_1.45 gridExtra_2.3
## [97] bookdown_0.37 xfun_0.41
## [99] stringi_1.8.3 yaml_2.3.8
## [101] evaluate_0.23 codetools_0.2-19
## [103] tibble_3.2.1 BiocManager_1.30.22
## [105] cli_3.6.2 xtable_1.8-4
## [107] munsell_0.5.0 jquerylib_0.1.4
## [109] survMisc_0.5.6 Rcpp_1.0.12
## [111] GenomicDataCommons_1.26.0 dbplyr_2.4.0
## [113] png_0.1-8 XML_3.99-0.16.1
## [115] ellipsis_0.3.2 readr_2.1.5
## [117] blob_1.2.4 prettyunits_1.2.0
## [119] bitops_1.0-7 scales_1.3.0
## [121] purrr_1.0.2 crayon_1.5.2
## [123] rlang_1.1.3 KEGGREST_1.42.0
## [125] rvest_1.0.3 formatR_1.14