This holds output from a call to zlm. Many methods are defined to operate on it. See below.

# S4 method for ZlmFit,CoefficientHypothesis
lrTest(object, hypothesis)

# S4 method for ZlmFit,Hypothesis
lrTest(object, hypothesis)

# S4 method for ZlmFit,matrix
lrTest(object, hypothesis)

# S4 method for ZlmFit,CoefficientHypothesis
waldTest(object, hypothesis)

# S4 method for ZlmFit,Hypothesis
waldTest(object, hypothesis)

# S4 method for ZlmFit
coef(object, which, ...)

# S4 method for ZlmFit
vcov(object, which, ...)

# S4 method for ZlmFit
se.coef(object, which, ...)

Arguments

object

ZlmFit

hypothesis

call to Hypothesis or CoefficientHypothesis or a matrix giving such contrasts.

which

character vector, one of "C" (continuous) or "D" (discrete) specifying which component should be returned

...

ignored

Value

see "Methods (by generic)"

Methods (by generic)

  • lrTest: Returns an array with likelihood-ratio tests on contrasts defined using CoefficientHypothesis().

  • lrTest: Returns an array with likelihood-ratio tests specified by Hypothesis, which is a Hypothesis.

  • lrTest: Returns an array with likelihood-ratio tests specified by Hypothesis, which is a contrast matrix.

  • waldTest: Returns an array with Wald Tests on contrasts defined using CoefficientHypothesis().

  • waldTest: Returns an array with Wald Tests on contrasts defined in Hypothesis()

  • coef: Returns the matrix of coefficients for component which.

  • vcov: Returns an array of variance/covariance matrices for component which.

  • se.coef: Returns a matrix of standard error estimates for coefficients on component which.

Slots

coefC

matrix of continuous coefficients

coefD

matrix of discrete coefficients

vcovC

array of variance/covariance matrices for coefficients

vcovD

array of variance/covariance matrices for coefficients

LMlike

the LmWrapper object used

sca

the SingleCellAssay object used

deviance

matrix of deviances

loglik

matrix of loglikelihoods

df.null

matrix of null (intercept only) degrees of freedom

df.resid

matrix of residual DOF

dispersion

matrix of dispersions (after shrinkage)

dispersionNoShrink

matrix of dispersion (before shrinkage)

priorDOF

shrinkage weight in terms of number of psuedo-obs

priorVar

shrinkage target

converged

output that may optionally be set by the underlying modeling function

hookOut

a list of length ngenes containing output from a hook function, if zlm was called with one

See also

zlm summary,ZlmFit-method

Examples

data(vbetaFA) zlmVbeta <- zlm(~ Stim.Condition+Population, subset(vbetaFA, ncells==1)[1:10,])
#> #> Done!
#Coefficients and standard errors coef(zlmVbeta, 'D')
#> (Intercept) Stim.ConditionUnstim PopulationCD154-VbetaUnresponsive #> B3GAT1 -5.4302447 1.00685352 0.90199793 #> BAX -0.4990626 0.02145208 -0.89154721 #> BCL2 -3.9407287 -2.20065589 -0.37551513 #> CCL2 -3.3547797 0.42749817 -0.29365929 #> CCL3 -2.6590780 -1.92367996 -0.28467650 #> CCL4 -3.3030403 -0.99943502 -0.93080019 #> CCL5 -1.8947363 0.11842660 0.23701817 #> CCR2 -3.7005996 1.30851899 0.02582989 #> CCR4 -0.3695476 -0.21384188 -0.43287571 #> CCR5 -2.8721780 -1.19817513 0.14714636 #> PopulationCD154+VbetaResponsive PopulationCD154+VbetaUnresponsive #> B3GAT1 1.6063715 -0.5811225 #> BAX -0.4386521 -0.5869687 #> BCL2 -0.3862382 1.3492612 #> CCL2 -0.3052702 -0.2626161 #> CCL3 0.4259786 -0.2617315 #> CCL4 0.5515850 0.3325102 #> CCL5 0.6887183 0.1555835 #> CCR2 -0.5950754 1.1230267 #> CCR4 0.2698394 -0.5880348 #> CCR5 0.1345247 0.6467222 #> PopulationVbetaResponsive PopulationVbetaUnresponsive #> B3GAT1 -1.2153975 0.322296844 #> BAX 0.1308370 0.111068557 #> BCL2 2.1094602 1.928661427 #> CCL2 -1.1332949 -2.305617623 #> CCL3 -0.4638945 -2.238025106 #> CCL4 -1.9149361 -1.918902719 #> CCL5 -0.1764152 -0.305926046 #> CCR2 -0.7973931 -0.150296597 #> CCR4 0.1166578 0.240411574 #> CCR5 -0.3888459 -0.005858029
coef(zlmVbeta, 'C')
#> (Intercept) Stim.ConditionUnstim PopulationCD154-VbetaUnresponsive #> B3GAT1 18.19441 -1.7969063 -0.55003066 #> BAX 17.47359 -0.6534516 -0.08181114 #> BCL2 16.41403 1.3286670 -0.86358539 #> CCL2 16.90490 -0.2937144 3.11986916 #> CCL3 19.81370 NA -0.55622889 #> CCL4 22.31976 NA -5.32517963 #> CCL5 19.22605 0.4504418 -0.68280704 #> CCR2 18.46733 -1.2269270 0.82209897 #> CCR4 17.45862 -0.3702826 -0.18436839 #> CCR5 15.77781 0.9461661 -0.15253610 #> PopulationCD154+VbetaResponsive PopulationCD154+VbetaUnresponsive #> B3GAT1 NA NA #> BAX 0.19649762 -0.03911775 #> BCL2 2.32150932 0.94613777 #> CCL2 7.04188443 3.34328706 #> CCL3 0.04812137 3.32313947 #> CCL4 -2.77190815 -1.51398583 #> CCL5 0.84757938 0.64493103 #> CCR2 -3.19304135 0.98294260 #> CCR4 0.57583680 0.47465344 #> CCR5 0.51211945 1.66591584 #> PopulationVbetaResponsive PopulationVbetaUnresponsive #> B3GAT1 NA NA #> BAX 0.1543621 0.1850906 #> BCL2 0.7362327 NA #> CCL2 NA NA #> CCL3 0.5075641 NA #> CCL4 NA NA #> CCL5 -0.7668751 -0.4885039 #> CCR2 0.1209953 -1.1165337 #> CCR4 0.3141728 -0.1774801 #> CCR5 -0.7502419 -1.0153526
se.coef(zlmVbeta, 'C')
#> #> X1 (Intercept) Stim.ConditionUnstim PopulationCD154-VbetaUnresponsive #> B3GAT1 1.1195139 1.9390550 1.9390550 #> BAX 0.2834781 0.2625570 0.3945849 #> BCL2 0.5300655 1.2983899 1.2983899 #> CCL2 2.5224874 4.3690763 3.5673358 #> CCL3 2.2877129 NA 3.0263597 #> CCL4 2.7577915 NA 4.7766350 #> CCL5 0.9651997 1.0038683 1.1357474 #> CCR2 1.1047933 1.2757054 1.3530899 #> CCR4 0.4658746 0.4372988 0.5966105 #> CCR5 1.0524207 2.1048414 1.2452410 #> #> X1 PopulationCD154+VbetaResponsive PopulationCD154+VbetaUnresponsive #> B3GAT1 NA NA #> BAX 0.3643741 0.4482183 #> BCL2 1.2983899 0.8655933 #> CCL2 3.5673358 4.3690763 #> CCL3 2.6825812 3.6171917 #> CCL4 3.2630629 3.9001061 #> CCL5 1.0847908 1.3068865 #> CCR2 1.5624136 1.2757054 #> CCR4 0.5532457 0.7298852 #> CCR5 1.2452410 1.2889469 #> #> X1 PopulationVbetaResponsive PopulationVbetaUnresponsive #> B3GAT1 NA NA #> BAX 0.3620264 0.3647365 #> BCL2 0.6940186 NA #> CCL2 NA NA #> CCL3 3.6171917 NA #> CCL4 NA NA #> CCL5 1.3267276 1.2838356 #> CCR2 1.8596442 1.5624136 #> CCR4 0.5968418 0.5927323 #> CCR5 1.8228462 1.3586693
#Test for a Population effect by dropping the whole term (a 5 degree of freedom test) lrTest(zlmVbeta, 'Population')
#> Refitting on reduced model...
#> #> Done!
#> , , metric = lambda #> #> test.type #> primerid cont disc hurdle #> B3GAT1 0.000000 4.299060 4.299060 #> BAX 1.072472 10.720978 11.793450 #> BCL2 5.443797 19.740650 25.184447 #> CCL2 4.103765 4.300998 8.404763 #> CCL3 1.523584 7.074064 8.597648 #> CCL4 1.516648 8.541585 10.058232 #> CCL5 5.572644 5.419661 10.992304 #> CCR2 12.540757 4.686256 17.227013 #> CCR4 4.607460 8.733661 13.341121 #> CCR5 7.136405 1.520790 8.657194 #> #> , , metric = df #> #> test.type #> primerid cont disc hurdle #> B3GAT1 0 5 5 #> BAX 5 5 10 #> BCL2 4 5 9 #> CCL2 3 5 8 #> CCL3 4 5 9 #> CCL4 3 5 8 #> CCL5 5 5 10 #> CCR2 5 5 10 #> CCR4 5 5 10 #> CCR5 5 5 10 #> #> , , metric = Pr(>Chisq) #> #> test.type #> primerid cont disc hurdle #> B3GAT1 1.00000000 0.507209744 0.50720974 #> BAX 0.95651117 0.057201590 0.29911799 #> BCL2 0.24471407 0.001397782 0.00277378 #> CCL2 0.25047521 0.506941990 0.39496253 #> CCL3 0.82245576 0.215190133 0.47521191 #> CCL4 0.67843337 0.128806114 0.26096254 #> CCL5 0.35004597 0.366835821 0.35811788 #> CCR2 0.02808429 0.455356085 0.06948928 #> CCR4 0.46563569 0.120170837 0.20521972 #> CCR5 0.21069179 0.910658274 0.56491639 #> #> attr(,"test") #> [1] "Population"
#Test only if the VbetaResponsive cells differ from the baseline group lrTest(zlmVbeta, CoefficientHypothesis('PopulationVbetaResponsive'))
#> Refitting on reduced model...
#> #> Done!
#> , , metric = lambda #> #> test.type #> primerid cont disc hurdle #> B3GAT1 0.000000000 0.49967559 0.4996756 #> BAX 0.187497834 0.05709950 0.2445973 #> BCL2 0.000000000 7.50139598 7.5013960 #> CCL2 0.000000000 1.51305236 1.5130524 #> CCL3 0.022203431 0.31780369 0.3400071 #> CCL4 0.000000000 2.44968166 2.4496817 #> CCL5 0.357112181 0.06828354 0.4253957 #> CCR2 0.006224572 0.58399389 0.5902185 #> CCR4 0.284055573 0.07376621 0.3578218 #> CCR5 0.210197743 0.11201241 0.3222102 #> #> , , metric = df #> #> test.type #> primerid cont disc hurdle #> B3GAT1 0 1 1 #> BAX 1 1 2 #> BCL2 0 1 1 #> CCL2 0 1 1 #> CCL3 1 1 2 #> CCL4 0 1 1 #> CCL5 1 1 2 #> CCR2 1 1 2 #> CCR4 1 1 2 #> CCR5 1 1 2 #> #> , , metric = Pr(>Chisq) #> #> test.type #> primerid cont disc hurdle #> B3GAT1 1.0000000 0.479642699 0.479642699 #> BAX 0.6650074 0.811140466 0.884884044 #> BCL2 1.0000000 0.006165119 0.006165119 #> CCL2 1.0000000 0.218673913 0.218673913 #> CCL3 0.8815472 0.572930548 0.843661814 #> CCL4 1.0000000 0.117548705 0.117548705 #> CCL5 0.5501144 0.793852536 0.808400350 #> CCR2 0.9371154 0.444751088 0.744450266 #> CCR4 0.5940553 0.785930097 0.836180408 #> CCR5 0.6466125 0.737864563 0.851202624 #> #> attr(,"test") #> [1] "PopulationVbetaResponsive"
# Test if there is a difference between CD154+/Unresponsive and CD154-/Unresponsive. # Note that because we parse the expression # the columns must be enclosed in backquotes # to protect the \quote{+} and \quote{-} characters. lrTest(zlmVbeta, Hypothesis('`PopulationCD154+VbetaUnresponsive` - `PopulationCD154-VbetaUnresponsive`'))
#> Warning: Some levels contain symbols. Be careful to escape these names with backticks ('`') when specifying contrasts.
#> Refitting on reduced model...
#> #> Done!
#> , , metric = lambda #> #> test.type #> primerid cont disc hurdle #> B3GAT1 0.000000000 0.53957307 0.5395731 #> BAX 0.009603320 0.41964445 0.4292478 #> BCL2 2.004979393 2.89187997 4.8968594 #> CCL2 0.003581828 -0.15725169 -0.1536699 #> CCL3 1.389919527 -0.60229580 0.7876237 #> CCL4 0.695474388 1.22512019 1.9205946 #> CCL5 1.643474587 0.02224792 1.6657225 #> CCR2 0.037354159 1.49555138 1.5329055 #> CCR4 0.977322744 0.14461070 1.1219334 #> CCR5 3.776168456 0.50471809 4.2808865 #> #> , , metric = df #> #> test.type #> primerid cont disc hurdle #> B3GAT1 0 1 1 #> BAX 1 1 2 #> BCL2 1 1 2 #> CCL2 1 1 2 #> CCL3 1 1 2 #> CCL4 1 1 2 #> CCL5 1 1 2 #> CCR2 1 1 2 #> CCR4 1 1 2 #> CCR5 1 1 2 #> #> , , metric = Pr(>Chisq) #> #> test.type #> primerid cont disc hurdle #> B3GAT1 1.00000000 0.46260972 0.4626097 #> BAX 0.92193505 0.51711451 0.8068449 #> BCL2 0.15678342 0.08902699 0.0864292 #> CCL2 0.95227640 1.00000000 1.0000000 #> CCL3 0.23841869 1.00000000 0.6744809 #> CCL4 0.40430856 0.26835815 0.3827791 #> CCL5 0.19984941 0.88142947 0.4348034 #> CCR2 0.84674577 0.22135713 0.4646584 #> CCR4 0.32286067 0.70373971 0.5706571 #> CCR5 0.05198758 0.47743434 0.1176027 #> #> attr(,"test") #> [1] "Contrast Matrix"
waldTest(zlmVbeta, Hypothesis('`PopulationCD154+VbetaUnresponsive` - `PopulationCD154-VbetaUnresponsive`'))
#> Warning: Some levels contain symbols. Be careful to escape these names with backticks ('`') when specifying contrasts.
#> , , metric = lambda #> #> test.type #> primerid cont disc hurdle #> B3GAT1 NA 0.4246764991 NA #> BAX 0.009305455 0.4499539395 0.459259394 #> BCL2 1.748463371 2.4975649351 4.246028306 #> CCL2 0.002614910 0.0007340968 0.003349006 #> CCL3 1.278016527 0.0007253470 1.278741874 #> CCL4 0.636616606 1.2610150758 1.897631682 #> CCL5 1.553679693 0.0235791990 1.577258892 #> CCR2 0.025434721 1.4461225520 1.471557273 #> CCR4 0.955370495 0.1340900313 1.089460527 #> CCR5 3.317281808 0.5214642628 3.838746071 #> #> , , metric = df #> #> test.type #> primerid cont disc hurdle #> B3GAT1 1 1 2 #> BAX 1 1 2 #> BCL2 1 1 2 #> CCL2 1 1 2 #> CCL3 1 1 2 #> CCL4 1 1 2 #> CCL5 1 1 2 #> CCR2 1 1 2 #> CCR4 1 1 2 #> CCR5 1 1 2 #> #> , , metric = Pr(>Chisq) #> #> test.type #> primerid cont disc hurdle #> B3GAT1 NA 0.5146127 NA #> BAX 0.92315144 0.5023568 0.7948279 #> BCL2 0.18607002 0.1140225 0.1196704 #> CCL2 0.95921700 0.9783846 0.9983269 #> CCL3 0.25826815 0.9785138 0.5276242 #> CCL4 0.42493864 0.2614590 0.3871993 #> CCL5 0.21259304 0.8779605 0.4544672 #> CCR2 0.87328861 0.2291511 0.4791322 #> CCR4 0.32835604 0.7142285 0.5799982 #> CCR5 0.06855509 0.4702177 0.1466989 #>