Concatenate specific lines and output the results in a new file

Question

I have multiple text files such as:

file1:

>ID_000_FLNNKGHD_01376 
-ATGAATACAGAGGAAAAAACACCGCTTGCATACAAT
>ID_000_MGCDKLCO_02388 
ATGAAGGTGGAAAAAACACCGCTTGCATTT
>ID_000_OMAMOGKP_02746
--ATGTTGGTGGAAAAAACACCGCTTGCGGTA

file2:

>ID_000_KGHDAAD_06245 
AAATACAGAGGAAAAAACACCGCTTGCATACAAT
>ID_000_KOAAFG_40481 
CCCCAGGAAGGTGGAAAAAACACCGCTTGCAAA
>ID_000_GPAAAGVV_07764
--AAATTGGTGG---------ACACCGCTTTT--

etc.

I need to concatenate the first and second lines, third and fourth lines, etc. of each file and output a new file for each result like this:

file1 output:

>ID_000_FLNNKGHD_01376-ID_000_KGHDAAD_06245 
-ATGAATACAGAGGAAAAAACACCGCTTGCATACAATAAATACAGAGGAAAAAACACCGCTTGCATACAAT

file2 output:

>ID_000_MGCDKLCO_02388-ID_000_KOAAFG_40481
ATGAAGGTGGAAAAAACACCGCTTGCATTTCCCCAGGAAGGTGGAAAAAACACCGCTTGCAAA

file3 output:

>ID_000_OMAMOGKP_02746-ID_000_GPAAAGVV_07764
--ATGTTGGTGGAAAAAACACCGCTTGCGGTA--AAATTGGTGG---------ACACCGCTTTT--

etc.

*note that the first line needs to start with ">" and is concatanated putting a "-" between each string, and the second line is just put together

outputs needs to be named file_1.fasta to file_n.fasta

Thanks!

Answer 1

Here's one way:

paste file1.fa file2.fa | 
    sed -E 's/\s+>/-/; s/\s+//g' | 
        awk -v c=0 '{ if(/^>/){c++} print > "file"c".pasted.fa"; }'

To explain this, let's have a look at what each command outputs:

$ paste file1.fa file2.fa 
>ID_000_FLNNKGHD_01376  >ID_000_KGHDAAD_06245 
-ATGAATACAGAGGAAAAAACACCGCTTGCATACAAT   AAATACAGAGGAAAAAACACCGCTTGCATACAAT
>ID_000_MGCDKLCO_02388  >ID_000_KOAAFG_40481 
ATGAAGGTGGAAAAAACACCGCTTGCATTT  CCCCAGGAAGGTGGAAAAAACACCGCTTGCAAA
>ID_000_OMAMOGKP_02746  >ID_000_GPAAAGVV_07764
--ATGTTGGTGGAAAAAACACCGCTTGCGGTA    --AAATTGGTGG---------ACACCGCTTTT--

So this will print each line from each file next to each other. Line 1 from file1 with line 1 from file 2, line 2 from file1 with line 2 from file2 etc. However, it has some extra spaces and an extra > which we need to get rid of. That's what the sed is doing:

$ paste file1.fa file2.fa | sed -E 's/\s+>/-/; s/\s+//' 
>ID_000_FLNNKGHD_01376-ID_000_KGHDAAD_06245
-ATGAATACAGAGGAAAAAACACCGCTTGCATACAATAAATACAGAGGAAAAAACACCGCTTGCATACAAT
>ID_000_MGCDKLCO_02388-ID_000_KOAAFG_40481
ATGAAGGTGGAAAAAACACCGCTTGCATTTCCCCAGGAAGGTGGAAAAAACACCGCTTGCAAA
>ID_000_OMAMOGKP_02746-ID_000_GPAAAGVV_07764
--ATGTTGGTGGAAAAAACACCGCTTGCGGTA--AAATTGGTGG---------ACACCGCTTTT--

The last step, the awk script will:

• awk -v c=0 : start awk and set the variable c to 0.
• if(/^>/){c++} : add 1 to the vallue of c every time we find a line that starts with >.
• print > "file"c".pasted.fa" : print the current line into a file called file, then the current value of c and the .pasted.fa.

The final result when run on your example is:

$ ls *pasted*
file1.pasted.fa  file2.pasted.fa  file3.pasted.fa

$ cat file1.pasted.fa 
>ID_000_FLNNKGHD_01376-ID_000_KGHDAAD_06245
-ATGAATACAGAGGAAAAAACACCGCTTGCATACAATAAATACAGAGGAAAAAACACCGCTTGCATACAAT
$ cat file2.pasted.fa 
>ID_000_MGCDKLCO_02388-ID_000_KOAAFG_40481
ATGAAGGTGGAAAAAACACCGCTTGCATTTCCCCAGGAAGGTGGAAAAAACACCGCTTGCAAA
$ cat file3.pasted.fa 
>ID_000_OMAMOGKP_02746-ID_000_GPAAAGVV_07764
--ATGTTGGTGGAAAAAACACCGCTTGCGGTA--AAATTGGTGG---------ACACCGCTTTT--