Question

我试图计算AATAA的100,000个核苷酸的5聚体中的5聚体的种群比例。数据中重复了多少个AATAA。

dog_ch38 <- read.GenBank("NC_006620.3")
dog_ch38 <- dog_ch38$NC_006620.3[1:100000]
dog_ch38 <- c2s(ape::as.character.DNAbin(dog_ch38))
dog_ch38 <- str_to_upper(dog_ch38)
 kmer_to_index <- function(kmer){
+     n  <- str_length(kmer)
+     letter_value  <- c("A" = 0, "C" = 1, "G" = 2, "T" = 3)
+     base  <- 1
+     index  <- 1
+     for( i in n:1){
+         nucleotide <- str_sub(kmer,start = i,end = i)
+         index  <- index + base * letter_value[nucleotide]
+         base  <- base * 4
+     }
+     return(as.numeric(index))
+ }
k <- 5
kmers <- numeric(4^k)
kmers
N <- str_length(dog_ch38)
> for (i in 1:(N - k + 1)) {
+   kmer <- str_sub(dog_ch38, i, i + k - 1)
+   index <- kmer_to_index(kmer)
+   kmers[index] <- kmers[index] + 1
+ }

但出现这样的错误：

Error in kmers[index] <- kmers[index] + 1 : 
  NAs are not allowed in subscripted assignments
In addition: Warning messages:
1: In 1:(N - k + 1) :
  numerical expression has 100000 elements: only the first used
2: In n:1 : numerical expression has 100000 elements: only the first used

seqinr::count(dog_ch38[1,], 5)
Error in dog_ch38[1, ] : incorrect number of dimensions

我实际上期望看到的结果是：

## aaaaa accaa taaag aataa
##   75   75     47    92

对于此功能，我绝对是新手，如果有人可以指导我如何解决它，请参考一些示例。谢谢！

Answer 1

我认为oligonucleotideFrequency()包中的Biostrings函数可以提供帮助。这是一个包含人工数据的示例。

  library(Biostrings) # requires appropriate Bioconductor install
  s1 <- sample(c("A", "C", "G", "T"), 10^5, TRUE)
  s1 <- DNAString(paste(s1, collapse = ""))
  kmers <- oligonucleotideFrequency(s1, width = 5)

该函数返回了所有可能 kmers的命名数字矢量。您可以使用此功能提取您感兴趣的kmer。在这个例子中应该有大约100个。

  kmers["AATAA"] # actual count varies because of random sampling
> AATAA 
>  102

查看此功能的帮助页面。使用默认参数，它将返回重叠 kmers。可以使用step选项来控制，如以下示例所示：

  s2 <- DNAString("AATAATAATAA")
  kmers1 <- oligonucleotideFrequency(s2, width = 5)
  kmers2 <- oligonucleotideFrequency(s2, width = 5, step = 5)

# See all the 5-mers found with step = 1 (default) versus step = 5
  kmers1[kmers1 != 0]
> AATAA ATAAT TAATA 
>     3     2     2 
  kmers2[kmers2 != 0]
> AATAA TAATA 
>     1     1

编辑

我发现（仍然发现）用于处理令人困惑的DNA序列的各种格式，并且似乎需要将read.GenBank()返回的紧凑型二进制格式转换为Biostrings中的字符表示形式。它们都非常有效。

可以在read.GenBank()返回的二进制对象列表上进行转换，也可以使用as.character = TRUE选项返回原始字符。我在这里展示后一种方法。

# Using package ape to read GenBank file, Biostrings for analysis
  library(ape)
  library(Biostrings)

# By default, read.GenBank returns a list of DNA sequences in compact binary form.
# This asks it to return a list of character vectors. 
  dog_ch38 <- read.GenBank("NC_006620.3", as.character = TRUE)
  str(dog_ch38)
> List of 1
>  $ NC_006620.3: chr [1:23914537] "n" "n" "n" "n" ...
>  - attr(*, "species")= chr "Canis_lupus_familiaris"

# Now convert the first (and only) member of the list to a single character string
  txt <- paste(dog_ch38[[1]], collapse = "")
  print(nchar(txt))
> [1] 23914537

# And now convert the character string to a DNAString
  s <- DNAString(txt)

# This is the form that can be handed to oligonucleotideFrequency
  km <- oligonucleotideFrequency(s[1:10^5], 5)
  km["AATAA"]
> AATAA 
>   176

如何计算kmer中的字母字符串

1 个答案:

编辑