richNull {GGtools} | R Documentation |
bind metadata concerning SNP allele frequency and other aspects of optimized cis-eQTL association to an mcwBestCis instance, to allow conditional FDR computation
richNull(..., MAFlb = 0.01, npc = 10, radius = 250000, nperm = 1, innerFilt = function(x) x, outerFilt = function(x) x) meta.richNull(..., MAFlb=.01, npc=10, radius=250000, nperm=1, innerFilt=function(x)x, outerFilt=function(x)x) # # internally: # # bigfilt = function(z) # outerFilt(MAFfilter(clipPCs(permEx(innerFilt(z)), 1:npc), lower=MAFlb)) #
... |
should provide bindings for smpack and chrnames,
which will be used to obtain gene/probe locations; see
meta.richNull allows a vector of smpack values bound to |
MAFlb |
lower bound on SNP MAF for null distribution evaluation |
npc |
number of expression principal components to be removed |
radius |
radius used for testing |
nperm |
This establishes how many permutations of expression against genotype will be performed for this process. |
innerFilt |
function immediately applied to generated smlSet instances |
outerFilt |
function applied to generated smlSet instances after clipPCs and MAFfilter are applied in that order |
The purpose of richNull
is to obtain realizations from the permutation
distribution of cis-eQTL association statistics, binding information on the
characteristics of the optimal results with the scores. This allows us to
use conditioning with the realizations from the permutation distribution.
richNull returns a list of nperm
mcwBestCis instances with
additional metadata bound in
Vince Carey <stvjc@channing.harvard.edu>