How to rename chromosome names in GTF file?

Question

[biomagician]

Hi,

I have a GTF file with the following head:

head celegans.gtf
CHROMOSOME_I    Coding_transcript   exon    4119    4358    .   -   transcript_id "Transcript:Y74C9A.3.1"; gene_id "Gene:Y74C9A.3";
CHROMOSOME_I    Coding_transcript   exon    5195    5296    .   -   transcript_id "Transcript:Y74C9A.3.1"; gene_id "Gene:Y74C9A.3";
CHROMOSOME_I    Coding_transcript   exon    6037    6327    .   -   transcript_id "Transcript:Y74C9A.3.1"; gene_id "Gene:Y74C9A.3";

However, my FASTA file has the following chromosome names:

grep '>' celegans.fa
>I
>II
>III
>IV
>V
>X
>MtDNA

This discrepancy causes problems in downstream analyses. Does anyone know of a tool or way to rename the chromosome names in my GTF file to correspond to the chromosome names in the FASTA file?

Thanks.

Best, C.

Answer 1

[cpad0112]

Try running sed -e 's/CHROMOSOME_//g' celegans.gtf. Check the output and if that works try : sed -i 's/CHROMOSOME_//g' celegans.gtf

[biomagician]

Hi,

sed -e 's/CHROMOSOME_//g' celegans.gtf

works but the command with the '-i' option gives the following error:

sed: 1: "output/genome/celegans/ ...": invalid command code o

Does the '-i' mean 'in-place' so changes the file directly? I am going to try to redirect the output of the '-e' command to the file itself.

Oups, this erased the content of the file. So the 'return' of the 'sed -e' command is NULL?

Best, C.

[cpad0112]

yes, unfortunately so. Before posting, I tried with example data and worked: (I am on Ubuntu and sed v4.2.2). I guess you are on MacOS and sed -i issue is discussed here and work around is given at the end of the post.

:~/Desktop $ sed -e 's/CHROMOSOME_//g' test.gtf 
head celegans.gtf
I    Coding_transcript   exon    4119    4358    .   -   transcript_id "Transcript:Y74C9A.3.1"; gene_id "Gene:Y74C9A.3";
I    Coding_transcript   exon    5195    5296    .   -   transcript_id "Transcript:Y74C9A.3.1"; gene_id "Gene:Y74C9A.3";
I    Coding_transcript   exon    6037    6327    .   -   transcript_id "Transcript:Y74C9A.3.1"; gene_id "Gene:Y74C9A.3";

:~/Desktop $ sed -i 's/CHROMOSOME_//g' test.gtf 

~/Desktop $ cat test.gtf 
head celegans.gtf
I    Coding_transcript   exon    4119    4358    .   -   transcript_id "Transcript:Y74C9A.3.1"; gene_id "Gene:Y74C9A.3";
I    Coding_transcript   exon    5195    5296    .   -   transcript_id "Transcript:Y74C9A.3.1"; gene_id "Gene:Y74C9A.3";
I    Coding_transcript   exon    6037    6327    .   -   transcript_id "Transcript:Y74C9A.3.1"; gene_id "Gene:Y74C9A.3";

$ uname -a
Linux genomics 4.8.0-53-generic #56~16.04.1-Ubuntu SMP Tue May 16 01:18:56 UTC 2017 x86_64 x86_64 x86_64 GNU/Linux

[biomagician]

Correct guess, thanks. It worked now. So you made use of the fact that the GTF file just had 'CHROMOSOME_' prepended to all my FASTA chromosome names, right? Do you mind explaining this: 's/CHROMOSOME_//g' ?

[cpad0112]

Correct. Sed syntax is s/old string /newstring/ (/ is a markup for before and after). g is for global replacement (entire file). Other wise only first match (old string) will be replaced. Entire substitution is in quotes. In above line, chromosome_ is old string and is replaced with no space in short it got removed.

[biomagician]

Maybe you can consolidate your comments into an answer and I can accept it?

[bruce.moran]

You could redirect to a new file and use that, no real need for -i

sed 's/CHROMOSOME_//g' celegans.gtf > celegans.noChr.gtf