Question

我有这两个数据集：之前包含5列（chromsome，start，end，行号，得分）

chrI         861         870    87     5
chrI         871         880    88    11
chrI         881         890    89    11
chrI         891         900    90    19
chrI         901         910    91    19
chrI         911         920    92    20
chrI         921         930    93    20
chrI         931         940    94    20
chrI         941         950    95    19
chrI         951         960    96    19
chrI         961         970    97    19
chrI         971         980    98    19
chrI         981         990    99    25
chrI         991        1000    100   20
chrI        1001        1010    101   20
chrI        1011        1020    102   20
chrI        1021        1030    103   20
chrI        1031        1040    104   15
chrI        1041        1050    105   14
chrI        1051        1060    106   14
chrI        1061        1070    107   13
chrI        1071        1080    108   13
chrI        1081        1090    109   13
chrI        1091        1100    110    7
chrI        1101        1110    111    7

峰值包含4列（chromsome，start，end，value）

"chrI"  880     1091    383
"chrI"  1350    1601    302
"chrI"  1680    1921    241
"chrI"  2220    2561    322
"chrI"  2750    2761    18
"chrI"  3100    3481    420
"chrI"  3660    4211    793
"chrI"  4480    4491    20
"chrI"  4710    4871    195
"chrI"  5010    5261    238

对于 Peaks 的每一行，我想在之前中提取相应的行（例如，第一行的880和1091之间的所有行），找到最高的得分值并将其写入新文件。的输出

  chrI       981         990    99    25

为此，我已经写了这个函数：

summit <- function(x,y,output){
    y<- Before
    chrom <- x[1]
    start <-x[2]
    end <-x[3]
    startLine <- y[which((y$V1 == chrom) & (y$V2==start)),]
    endLine <- y[which((y$V1 == chrom) & (y$V3==end)),]
    Subset <- y[which((y$V2 >= startLine$V2) & (y$V3 <= endLine$V2))]
    maximum <- Subset[which(Subset$V4 == max(Subset$V4))]
    output <- print(maximum)
}

apply(Peaks,1,summit,output = 'peaks_list.bed')

我没有收到错误消息，但它在整个晚上都没有给我结果，所以我猜我的代码有问题，但我不知道是什么。

我也试试这个：

Peaks_Range <- GRanges(seqnames=Peaks$V1, ranges=IRanges(start=Peaks$V2, end=Peaks$V3))
Before_Range <- GRanges(seqnames=Before$V1, ranges=IRanges(start=Before$V2, end=Before$V3),score=Before$V5)
Merged <- mergeByOverlaps(Peaks_Range,Before_Range)
Merged <- as.data.frame(Merged)


for (i in 1:nrow(Peaks)){
        start <-Peaks[i,2]
        end <-Peaks[i,3]
        Subset <- subset(Merged,Merged$Peaks_Range.start == start)
        maximum <- as.numeric(max(Subset$score))
        summit <- Subset[which(Subset$score == maximum),]
        write.table(summit,'peaks_list.bed', sep="\t", append=T, col.name = FALSE, row.names = FALSE, quote = FALSE)

}

它有效（我认为），但这非常慢，所以我搜索另一种方法来做到这一点。

有没有人有任何想法？

Answer 1

您可以使用cut来帮助您获取索引。

setwd("/home/wang/Downloads")

before <- read.table("before.txt", header = F, stringsAsFactors = F)
colnames(before) <- c("chromosome", "start", "end", "line number", "score")
peaks <- read.table("peaks.txt", header = F, stringsAsFactors = F, quote = "\"")
colnames(peaks) <- c("chromosome", "start", "end", "value")

summit <- function(peaks_vec){
  chromosome = peaks_vec[1]
  start = as.numeric(peaks_vec[2])
  end = as.numeric(peaks_vec[3])
  filter_before = subset(before, chromosome == chromosome)
  up_index = cut(end, filter_before[,2], labels = F) +1
  down_index = cut(start, filter_before[,2], labels = F) +1
  if(!is.na(down_index) & !is.na(up_index)){
    new_filter_before = filter_before[down_index : up_index, ]
    max_index = which.max(new_filter_before[,5])
    return(unlist(new_filter_before[max_index,]))
  }else {
    return(rep(NA, 5)) # you can input what you want.
  }
}
result <- t(apply(as.matrix(peaks), 1, summit))
remove_na_result <- as.data.frame(na.omit(result))
colnames(remove_na_result) <- colnames(before)

最终输出是：

  chromsome start end line number score
1      chrI   981 990          99    25

希望我的回答很有帮助。

在另一个文件定义的间隔中查找最高值

1 个答案: