UniD.probefilter: Probe and sample filter

In JieYang031/UniD: Unified Diagnostic Platform for Gliomas

Description

Used to filter out probes with suspect quality issues for following categories: (1) probes located on ChrX/Y; (2) probes may affected by SNP; (3) probes may mapped to multiple locations; (4) probes not targeted CpG sites; (5) probes on 450k platform but not available on EPIC platform; (6) probes with large proportion of missing values. It can also filter out samples with large proportion of missing values. Missing values may caused by non-significant detection p-value or less beadcount per probe.

Usage

UniD.probefilter(Beta.raw, outDir, filterXY = TRUE, filterSNPHit = TRUE,
  filterMultiHit = TRUE, filterNonCG = FALSE, filterNonEpic = TRUE,
  arrayType = c("450k", "EPIC"), filterSample = TRUE,
  filterSample.cut = 0.1, filterProbe = FALSE, filterProbe.cut = 0.05,
  write)

Arguments

Beta.raw	data frame generated from the UniD.dataqc()
outDir	directory where output data should be saved if write = T
filterXY	whether filter out probes located on Chromosome X and Y. Default value is TRUE.
filterSNPHit	whether filter out probes may affected by SNP. Default is TRUE. Probe list adapted from Zhou W, Nucleic Acids Research, 2017.
filterMultiHit	whether filter out probes can mapped to multi locations. Default is TRUE. Probe list adapted from Nordlund J, Genome Biology, 2013.
filterNonCG	whether filter out probes which are not targeting CpG sites. Default is FALSE.
filterNonEpic	whether filter outprobes which are available on 450k platform but not available on EPIC platform. Default is T. Highly recommended if building models with data from 450k platform.
arrayType	the platform which raw data generated, can be "450k" or "EPIC"
filterSample	whether filter out samples with high proportion of missing values. Default is TRUE.
filterSample.cut	if filterSample = T, the threshold for high proportion of missing values per sample. Default is 0.1.
filterProbe	whether filter out probes with high proportion of missing values. Default is TRUE. Carefully usage with small number of sample size.
filterProbe.cut	If filterProbe = T, the threshold for high proporition of missing values per probe. Default is 0.05.
write	whether the output should be saved, highly recommended

Value

A data frame with probes after filtering in Beta value filtering

References

Zhou, W., et al. (2017). "Comprehensive characterization, annotation and innovative use of Infinium DNA #' methylation BeadChip probes." Nucleic Acids Res 45(4): e22.

Nordlund, J., et al. (2013). "Genome-wide signatures of differential DNA methylation in pediatric acute #' lymphoblastic leukemia." Genome Biol 14(9): r105.

Examples

## Not run: 
Beta.clean <- UniD.probefilter(Beta.raw, outDir = "~/Desktop/output/",
filterXY = T, filterSNPHit = T, filterMultiHit = T, filterNonCG = F,
filterNonEpic = T, arrayType = "450k", filterSample = T, filterSample.cut
= 0.1, filterProbe = F, write = T)

Beta.clean <- UniD.probefilter(Beta.raw, outDir = NULL,
filterXY = T, filterSNPHit = T, filterMultiHit = T, filterNonCG = F,
filterNonEpic = F, arrayType = "EPIC", filterSample = T, filterSample.cut
= 0.1, filterProbe = F, write = F)

## End(Not run)

JieYang031/UniD documentation built on May 5, 2021, 5:16 p.m.