Combining two fasta sequences into one

Question

[Lille My]

I have two fasta files, with the same headers/names for the sequences but different sequences. I would like to combine them into one file, so that each sequence has the same name but is a combination of both sequences. My preferred language is bash script, but I'm open to other suggestions. thanks.

[Pierre Lindenbaum]

with the same headers/names for the sequences but different sequences

uhh ?

would like to combine them into one file, s

an example is needed

[GenoMax]

Like this?

File_1:

>Seq_1
ACGCTAGCTA
>Seq_2
CGCTAGCTC

File_2:

>Seq_1
GCTGAT
>Seq_2
TTACTC

File_1 + File_2 = File_3

>Seq_1
ACGCTAGCTAGCTGAT
>Seq_2
CGCTAGCTCTTACTC

[Lille My]

yes, your example is exactly what I need to do.

[WouterDeCoster]

Does this make biological sense?

[Lille My]

Sometimes it does, depends on what kind of sequences you have.

Answer 1

[shenwei356]

A solution using seqkit, csvtk and shell sed.

Sample files (not in same order, can be multiple lines):

$ cat 1.fa
>seq1
aaa
aa
>seq2
ccc
cc
>seq3
ggg
gg

$ cat 2.fa
>seq3
TTT
TT
>seq2
GGG
GG
>seq1
CCC
CC

Just one command:

$ seqkit concat 1.fa 2.fa
>seq1
aaaaaCCCCC
>seq2
cccccGGGGG
>seq3
gggggTTTTT

Step 1. Convert FASTA to tab-delimited (3 columns, the 3rd column is blank (no quality for FASTA)) file:

$ seqkit fx2tab 1.fa > 1.fa.tsv
$ seqkit fx2tab 2.fa > 2.fa.tsv

$ cat -A 1.fa.tsv 
seq1^Iaaaaa^I$
seq2^Iccccc^I$
seq3^Iggggg^I$

Step 2. Merge two table files:

$ csvtk join -H -t 1.fa.tsv 2.fa.tsv | cat -A
seq1^Iaaaaa^I^ICCCCC^I$
seq2^Iccccc^I^IGGGGG^I$
seq3^Iggggg^I^ITTTTT^I$

Step 3. Note that there are two TAB between the two sequences, so we can remove them to join the sequences

$ csvtk join -H -t 1.fa.tsv 2.fa.tsv | sed 's/\t\t//'
seq1    aaaaaCCCCC
seq2    cccccGGGGG
seq3    gggggTTTTT

Step 4. Convert tab-delimited file back to FASTA file:

$ csvtk join -H -t 1.fa.tsv 2.fa.tsv | sed 's/\t\t//' | seqkit tab2fx
>seq1
aaaaaCCCCC
>seq2
cccccGGGGG
>seq3
gggggTTTTT

All in one command:

$ csvtk join -H -t <(seqkit fx2tab 1.fa) <(seqkit fx2tab 2.fa) | sed 's/\t\t//' | seqkit tab2fx

[Lille My]

thanks! I will try this out.

Answer 2

[Pierre Lindenbaum]

assuming there are only twho lines per sequence (title/dna) and they are ordered the same way.

paste  f1.fa f2.fa | sed -e 's/\t>.*//' -e 's/\t//'

[Lille My]

there are more lines, but I can change them into one liners. I will try this out. thanks!