intensity.vs.ref: KS-test for difference in intensity distributions

In andrewparkermorgan/argyle: Basic interface and QC for genotypes from Illumina Infinium arrays

Description

KS-test for difference in intensity distributions

Usage

## S3 method for class 'vs.ref'
intensity(gty, ref, ...)

Arguments

gty	a genotypes object with intensity data attached
ref	a matrix, vector or object coercible to such, containing sum-intensities from reference samples
...	ignored

Details

This function detects potential failed arrays by performing the Kolmogorov-Smirnov test for difference between the "sum-intensities" (see summarize.intensity) of each sample and some reference distribution of "sum-intensities" of known good samples. This test should (obviously) be performed *before* any normalizations are applied. As such it may be useful for detecting batch effects, although that possibility has not been systematically explored.

The distribution of "sum-intensity" across an array is expected to be approximately normal. Outliers for the D statistic come in two flavours (cf. Didion et al. (2014)): samples which fail completely, having a heavily right-skewed intensity distribution; and samples which are genetically diverged from the reference sample/species used in array design. Divergent samples have a spike of intensities near zero, representing failed hybridization due to off-target variation, but an otherwise normal intensity distribution.

Value

a named vector of D_j, the Kolmogorov-Smirnov test statistic for each sample j

References

Didion JP et al. (2014) SNP array profiling of mouse cell lines identifies their strains of origin and reveals cross-contamination and widespread aneuploidy. BMC Genomics 15(1): 847. doi:10.1186/1471-2164-15-847.

See Also

summarize.intensity, summarize.calls, run.sample.qc

andrewparkermorgan/argyle documentation built on May 10, 2019, 11:08 a.m.