数据选择错误

时间:2013-09-16 06:46:13

标签: r

当我尝试根据小数据集(d2)从大型数据集(d1)中选择数据时出错。下面是我的脚本和问题。

**d1 <- read.table("MSv25.txt",header=T)

d2 <- read.table("Flairall.Gene.txt",header=T)

d2$low <- d2$start-10000 ; d2$high <- d2$end+10000

d1$matched <- apply(d1,1,function(p) which(p['POS'] >=d2[,'low'] & p['POS'] <= d2[,'high'] & p['CHR']==d2[,"chromosome"]))

d3 <- cbind(d1[which(d1$matched >0),], d2[unlist(d1$matched[which(d1$matched>0)]),])

write.table(d3,file="Flairall.GOBSgene.txt",quote=FALSE,sep="\t",row.names=FALSE,col.names=TRUE)**

d1是这样的:

SNP CHR POS A1  A2  OR  P
rs1000007   2   237416793   C   T   0.9785  0.4868
rs1000003   3   99825597    G   A   0.9091  0.009774
rs1000002   3   185118462   C   T   1.0111  0.6765
rs10000012  4   1347325 G   C   1.0045  0.9087
rs10000042  4   5288053 T   C   1.0622  0.3921
rs10000062  4   5305645 G   C   1.0116  0.779
rs10000132  4   7450570 T   C   0.9734  0.4748
rs10000081  4   16957461    C   T   1.0288  0.3585
rs10000100  4   19119591    A   G   1.0839  0.1417
rs10000010  4   21227772    C   T   0.971   0.2693
rs10000092  4   21504615    C   T   1.0342  0.27
rs10000068  4   36600682    T   C   1.055   0.103
rs10000013  4   36901464    C   A   1.0198  0.5379
rs10000037  4   38600725    A   G   1.0249  0.4217
rs10000017  4   84997149    T   C   0.9576  0.1912
rs10000109  4   91586292    A   T   0.9956  0.9349
rs10000023  4   95952929    T   G   0.9998  0.9951
rs10000030  4   103593179   A   G   1.0839  0.04208
rs10000111  4   107137517   A   G   1.0812  0.3128
rs10000124  4   109325900   A   C   1.0642  0.1906
rs10000064  4   128029071   C   T   1.0388  0.1578
rs10000029  4   138905074   C   T   0.7832  0.14
rs10000036  4   139438712   T   C   0.9848  0.5683
rs10000033  4   139819348   C   T   0.9918  0.7542
rs10000121  4   157793485   A   G   1.0008  0.9769
rs10000041  4   165841405   G   T   1.0042  0.9146
rs10000082  4   167529642   T   C   0.9733  0.6612

d2是这样的:

gene    start   end chromosome
WFDC9   237416000   237418000   2
SRGAP3  19119590    21504615    4

一般来说,我想通过在开始和结束位置延伸10kb的窗口来选择基因内的SNP。

这是我的脚本结果:

   SNP CHR       POS A1 A2     OR      P matched  gene     start       end chromosome       low      high
    1 rs1000007   2 237416793  C  T 0.9785 0.4868       1 WFDC9 237416000 237418000          2 237406000 237428000

哪个不正确,因为缺少一个基因。正确的应该是:

gene    start   end chromosome  SNP CHR POS A1  A2  OR  P
WFDC9   237416000   237418000   2   rs1000007   2   237416793   C   T   0.9785  0.4868
SRGAP3  19119590    21504615    4   rs10000100  4   19119591    A   G   1.0839  0.1417
SRGAP3  19119590    21504615    4   rs10000010  4   21227772    C   T   0.971   0.2693
SRGAP3  19119590    21504615    4   rs10000092  4   21504615    C   T   1.0342  0.27

任何人都可以帮我指出什么是错的......非常感谢......

1 个答案:

答案 0 :(得分:2)

您的代码似乎完全落后于您要实现的目标:

“对于每个基因(在d2中)哪个SNP(来自d1)在该基因的10kb内?”

首先,d1$matched的代码是向后的。你的所有pd2都应该是相反的(目前它没有多大意义?),给你一个与每个基因顺式相关的SNP列表(+ / - 10kb)。

我会按照我提出问题的方式接近它:

cisWindow <- 10000 # size of your +/- window, in this case 10kb.
d3 <- data.frame()
# For each gene, locate the cis-SNPs
for (i in 1:nrow(d2)) {
  # Broken down into steps for readability.
  inCis <- d1[which(d1[,"CHR"] == d2[i, "chromosome"]),]
  inCis <- inCis[which(inCis[,"POS"] >= (d2[i, "start"] - cisWindow)),]
  inCis <- inCis[which(inCis[,"POS"] <= (d2[i, "end"] + cisWindow)),]
  # Now we have the cis-SNPs, so lets build the data.frame for this gene,
  # and grow our data.frame d3:
  if (nrow(inCis) > 0) {
    d3 <- rbind(d3, cbind(d2[i,], inCis))
  }
}

我试图找到一个不涉及在循环中增长d3的解决方案,但因为您将d2的每一行附加到d1的0行或更多行中无法提出一个效率不高的解决方案。