マッチングとフィルタリング

Rで始めるBioconductor入門

Paula Andrea Martinez, PhD.

Data Scientist

重複配列

生物学的配列の重複は自然界でも発生する
ライブラリ調製工程（PCR）による増幅
サンプルの複数回シーケンシング

重複の除去またはマーキング

ゲノム全域・エクソームシーケンシング

閾値を用いた重複のマーキング

RNA-seq・ChIP-seq

srduplicated

library(ShortRead)

# Counting duplicates TRUE is the number of duplicates
table(srduplicated(dfqsample))

FALSE  TRUE 
500   500

# Cleaning reads from duplicates x[fun(x)]
cleanReads <- mydReads[srduplicated(mydReads) == FALSE]

# Counting duplicates
table(srduplicated(cleanReads))

FALSE
500

カスタムフィルターの作成

srFilter で条件に基づくフィルタリング x[fun(x)]

フィルターの例

library(ShortRead)

# Use a custom filter to remove reads from fqsample
# This filter to remove reads shorter than a min number of bases
readWidthCutOff <- srFilter(function(x) {width(x) >= minWidth}, 
                            name = "MinWidth")

minWidth <- 51

fqsample[readWidthCutOff(fqsample)]

nFilter

library(ShortRead)

# save your filter, .name is optional
myFilter <- nFilter(threshold = 10, .name = "cleanNFilter")

# use the filter at reading point    
filtered <- readFastq(dirPath = "data", 
                      pattern = ".fastq",
                       filter = myFilter)

# you will retrieve only those reads that have a maximum of 10 N's
filtered

idFilter と polynFilter

library(ShortRead)

#id filter example
myFilterID <- idFilter(regex = ":3:1")
# will return only those ids that contain the regular expression

# optional parameters are .name, fixed and exclude
# use the filter at reading point    
filtered <- readFastq(dirPath = "data", pattern = ".fastq",
                       filter = myFilterID)

# filter to remove poly-A regions
myFilterPolyA <- polynFilter(threshold = 10, nuc = c("A"))
# will return the sequences that have a maximun number of 10 consecutive A's

# use the filter for subsetting    
filtered[myFilterPolyA(filtered)]

フィルターを使って練習しましょう！

Rで始めるBioconductor入門