filter_probes.Rmd

---
title: "Filter probes"
author: "The methylation in ART meta-analysis group"
date: "`r Sys.Date()`"
output: html_document
---
## Introduction

It is well known that some infinium probes have issues, such as overlapping SNPs, non-unique mapping or other problems.
While analysing a recent dataset (Estill 2016), I saw that some of the methylation data appeared to be affected by SNPs, with distinct populations of normal/heterozyzous/homozygous variant.
While this might be interesting from a genetic perspective, we should be filtering out these probes for methylation analysis.
In this short workflow, I will use the advice from Zhou et al (2017) to create a list of probes that should be retained.
While designed for use with EPIC arrays, this should also work for 450k array because most good probed in the 450k design are retained for EPIC.

The list of good probes is called "good_probes.txt".

The list of bad probes is called "bad_probes.txt".

## Download data

In the chunk here I'm downloading the supplementary data from the Zhou et al (2017)

```{r dl}
dir.create("filter_probes")

download.file("https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5389466/bin/gkw967_supplementary_data.zip",destfile = "filter_probes/gkw967_supplementary_data.zip")

unzip(zipfile = "filter_probes/gkw967_supplementary_data.zip", exdir = "filter_probes")

unzip(zipfile = "filter_probes/nar-01910-met-k-2016-File009.zip",exdir = "filter_probes")

x <- read.table("filter_probes/EPIC.anno.GRCh38.tsv", sep="\t", header=TRUE,stringsAsFactors = FALSE)
head(x)
names(x)
dim(x)
head(x$MASK.general)

bad <- x[which(x$MASK.general==TRUE),"probeID"]
head(bad)
length(bad)
write.table(bad,file="probes_bad.txt")
write.table(bad,file="filter_probes/probes_bad.txt")

good <- x[which(x$MASK.general==FALSE),"probeID"]
head(good)
length(good)
write.table(good,file="probes_good.txt")
write.table(good,file="filter_probes/probes_good.txt")

```

## Session info

```{r,sessioninfo}
sessionInfo()
```

## References
Zhou W, Laird PW, Shen H. Comprehensive characterization, annotation and innovative use of Infinium DNA methylation BeadChip probes. Nucleic Acids Res. 2017;45(4):e22. doi:10.1093/nar/gkw967