如何从重叠的基因列表中提取基因名称？

Question

我目前有一个脚本，研究不同基因列表之间的基因重叠。因此，代码返回的是一个8乘8的矩阵，其中包含发现重叠的不同数量的基因(即两个列表中共有的基因)。有没有一种方法可以让我观察这些特定的基因并找出它们的基因符号？

#-------------------------------------------------------------------------------
# Set the working directory and load the data files
#-------------------------------------------------------------------------------
setwd("~/Desktop/R_Project/Gene_overlap")
getwd()
files <- list.files(pattern="*.txt", full.names = TRUE)
files

data.list <- lapply(files, function(fil) {
  scan(file=fil, what=character())
})

names(data.list) <- basename(files) %>% stringr::str_remove("\\.txt$")

str(data.list)
# List of 8
# $ GSE108363_BCGdown_D:chr [1:350] "IL1B" "IL6" "IL1A" "CCL20" ...
# $ GSE108363_BCGdown_V: chr [1:267] "IL6" "CCL20" "IL1A" "CXCL5" ...
# $ GSE108363_BCGup_D  : chr [1:250] "FABP4" "CMTM2" "FUCA1" "CD36" ...
# $ GSE108363_BCGup_V  : chr [1:429] "FCN1" "FCGR3B" "MNDA" "CPVL" ...
# $ GSE108363_MTBdown_D: chr [1:86] "CCL20" "IL1B" "IL1A" "IL6" ...
# $ GSE108363_MTBdown_V: chr [1:244] "IL1B" "IL1A" "CCL20" "IL6" ...
# $ GSE108363_MTBup_D  : chr [1:128] "FUCA1" "FGL2" "TGFBI" "CPVL" ...
# $ GSE108363_MTBup_V  : chr [1:286] "FABP4" "RNASE1" "MNDA" "CPVL" ...

intersect(data.list$GSE108363_BCGdown_D, data.list$GSE108363_BCGdown_V) %>% length

sapply(data.list, length)



#-------------------------------------------------------------------------------
# Using the intersect function to see the overlaps 
#-------------------------------------------------------------------------------

data.file1 <- "GSE108363_BCGdown_D.txt"
data.file2 <- "GSE108363_BCGdown_V.txt"
data.file3 <- "GSE108363_BCGup_D.txt"
data.file4 <- "GSE108363_BCGup_V.txt"
data.file5 <- "GSE108363_MTBdown_D.txt"
data.file6 <- "GSE108363_MTBdown_V.txt"
data.file7 <- "GSE108363_MTBup_D.txt"
data.file8 <- "GSE108363_MTBup_V.txt"

genevect1 <- scan(data.file1, what=character(), sep="\n")
genevect2 <- scan(data.file2, what=character(), sep="\n")
genevect3 <- scan(data.file3, what=character(), sep="\n")
genevect4 <- scan(data.file4, what=character(), sep="\n")
genevect5 <- scan(data.file5, what=character(), sep="\n")
genevect6 <- scan(data.file6, what=character(), sep="\n")
genevect7 <- scan(data.file7, what=character(), sep="\n")
genevect8 <- scan(data.file8, what=character(), sep="\n")


filelist <- list(data.file1, data.file2, data.file3, data.file4, data.file5, data.file6, data.file7, data.file8)
all(sapply(filelist, file.exists))

#-------------------------------------------------------------------------------
# read files:
#-------------------------------------------------------------------------------

gene.lists <- lapply(filelist, function(f) {
  scan(file=f, what=character())
})


#-------------------------------------------------------------------------------
# set up empty matrix for overlaps
#-------------------------------------------------------------------------------
x <- (length(gene.lists))^2
x
y <- rep(NA, x)
mx <- matrix(y, ncol=length(gene.lists))
mx
row.names(mx) <- sapply(filelist, basename) %>% stringr::str_remove('.txt$')
colnames(mx) <- sapply(filelist, basename) %>% stringr::str_remove('.txt$')
mx

mx.overlap.count <- mx

#-------------------------------------------------------------------------------
# Overlaps
#-------------------------------------------------------------------------------
for (i in seq_along(gene.lists)) {
  g1 <- gene.lists[[i]]
  for (j in seq_along(gene.lists)) {
    g2 <- gene.lists[[j]]
    a <- intersect(g1, g2)
    b <- length(a)
    mx.overlap.count[j,i] <- b
  }
}

mx.overlap.count
round(as.numeric(mx.overlap.count),digits = 1)
View(mx.overlap.count)

目前，此代码返回数值。然而，我想为每个被比较的两个基因列表生成某种类型的列表(或者类似的东西)，这样我就可以确切地看到哪些基因是两者共同的。

Answer 1

我会通过循环对比矩阵来解决这个问题，像您一样将交集的长度存储在矩阵中，并将基因名称存储在列表genes.overlap中。如下所示：

# Load files. 
file_names <- list.files(pattern=".txt")

# Extract gene lists. 
gene.lists <- lapply(file_names, function(f) {
  scan(file=f, what=character())
})

# Name the entries in the list. 
names(gene.lists) <- file_names
names(gene.lists)

# Initiate an empty list and matrix for storing output of loop.
genes.overlap <- list()
nfiles <- length(gene.lists)
mx.overlap.count <- matrix(NA,nrow=nfiles)

# Generate contrasts:
contrasts <- combn(nfiles,2)

# Loop to determine intersection:
for (i in 1:dim(contrasts)[2]){
  list1 <- contrasts[1,i]
  list2 <- contrasts[2,i]
  g1 <- gene.lists[[list1]]
  g2 <- gene.lists[[list2]]
  comparison_name <- paste(names(gene.lists[list1]),names(gene.lists[list2]),sep="_")
  genes.overlap[[i]] <- intersect(g1, g2)
  names(genes.overlap)[i] <- comparison_name
  b <- length(genes.overlap[[i]])
  mx.overlap.count[i] <- b
}

# You can index into a list like a df with the $ operator. 
genes.overlap$List.txt_List1.txt