关于最短子串实现问题，如何提高R代码的速度？

编程问答 2022-06-11

问题描述

我决定在R中实施Rosalind的this问题。它适用于小型样本数据集：

>Rosalind_56
ATTAGACCTG
>Rosalind_57
CCTGCCGGAA
>Rosalind_58
AGACCTGCCG
>Rosalind_59
GCCGGAATAC

但是，当我为真实数据集（50个序列，长度约为1000个字符）运行它时，会花费很长时间。 您能否看一下我的代码并提供一些建议，我应该替换一下以提高速度？我只有五分钟的时间来发布答案...这是我写的一些解释。

library(Biostrings)
library(stringr)

fastaFile = readDNAStringSet("input.fasta")
seq_name = names(fastaFile)
sequence = paste(fastaFile)
df <- data.frame(seq_name,sequence) #read fasta and convert it to simple df

allsubstr <- function(x,n) unique(substring(x,1:(nchar(x) - n + 1),n:nchar(x))) #helping function,creating all substring of given string

Shortest_substring_R <- function(df){
  
tmp_ss <- df[1,2] #first element will be modified to final result
tmp_df <- df #put everything to temporary df which will be also modified (number of elements)

while(nrow(tmp_df)>1){ #loop until we heave more than one sequence in our temp df
max_ov_all <- '' #it's better to clear it
max_ov <- ''
for(i in 2:nrow(tmp_df)){ #this loop finds the maximum overlap between result and ith sequence (result - shortest substring)
res <- unlist(sapply(1:nchar(tmp_df[i,2]),function(x) allsubstr(tmp_df[i,2],x))) #produce all substrings from ith sequence
ovs <- res[str_detect(tmp_ss,res)] #only substrings of ith sequence which overlaps with first (result) sequence
max_ov <- ovs[which.max(nchar(ovs))] #give me the longest common substring of first (result) and ith sequence
if(nchar(max_ov) > nchar(max_ov_all)){ #check which overlap is the longest (between first (result) and remaining sequences)
  max_ov_all <- max_ov
  ind <- i
}
}
if(str_detect(tmp_ss,tmp_df[ind,2])){ #Now three cases,our substring (coming from ith sequence) is inside our result string
  tmp_df <- tmp_df[-ind,] #so we do not have do anything,juest remove it from the df
}else{ #our first (result) sequence may start or end with longest overlap,so modify final result as needed
       #and then of course remove ith sequence (giving longest overlap),because it has been added to aour result sequence
  if(startsWith(tmp_ss,max_ov_all)){
    tmp_ss <- paste(tmp_df[ind,sub(max_ov_all,"",tmp_ss),sep='')
    tmp_df <- tmp_df[-ind,]
  }
  if(endsWith(tmp_ss,max_ov_all)){
    tmp_ss <- paste(sub(max_ov_all,]
  }
}
}
print(tmp_ss)
}

结果：

> Shortest_substring_R(df)
[1] "ATTAGACCTGCCGGAATAC"

我认为主要问题在于for()循环，sapply()和非常有用的功能，例如str_detect()您怎么看？

样本输入：

Input = ('seq_name   sequence
1 Rosalind_56 ATTAGACCTG
2 Rosalind_57 CCTGCCGGAA
3 Rosalind_58 AGACCTGCCG
4 Rosalind_59 GCCGGAATAC')
small_df = as.data.frame(read.table(textConnection(Input),header = T,row.names = 1))
small_df

解决方法

暂无找到可以解决该程序问题的有效方法，小编努力寻找整理中！

如果你已经找到好的解决方法，欢迎将解决方案带上本链接一起发送给小编。

小编邮箱:dio#foxmail.com (将#修改为@）

dataframerrrosalindsapplystringstring