kolmogorov_batches: Group provisional batch labels by similarity of eCDFs

In scristia/CNPBayes: Bayesian mixture models for copy number polymorphisms

Description

This function groups provisional batches (e.g., chemistry plate, date, or DNA source) by similarity of the empirical cummulative distribution function (eCDF). Similarity of the eCDFs is based on the p-value from Kolmogorov-Smirnov test statistic. All pairwise combinations of batches are compared recursively until no two batches can be combined.

Usage

kolmogorov_batches(dat, KS_cutoff)

Arguments

dat	typically a 'tibble' gotten by 'assays(MultiBatch)'
KS_cutoff	scalar indicating cutoff for Kolmogorov-Smirnov p-value. Two provisional batches with p-value above this cutoff are combined.

See Also

ks.test

Examples

extdir <- system.file("extdata", package="CNPBayes")
se <- readRDS(file.path(extdir, "snp_se.rds"))
cnv_region <- GRanges("chr2", IRanges(90010895, 90248037),
                      seqinfo=seqinfo(se))
se2 <- subsetByOverlaps(se, cnv_region)
provisional_batch <- se2$Sample.Plate
full.data <- median_summary(se2,
                            provisional_batch=provisional_batch,
                            assay_index=2,
                            THR=-1)
## Not run: 
   batched.data <- kolmogorov_batches(full.data, 1e-6)

## End(Not run)

scristia/CNPBayes documentation built on Aug. 9, 2020, 7:31 p.m.