Extracting only the first match from FASTA file

[Henry_C]

Hello to everybody.
I am a plant biologist trying to switch to bioinformatics (notably, RNA-Seq). It is just a few months agoi I discovered all the Linux world and its surroundings. Now that I am going on with the analysis, I do find more and more bio-informatics problems that probably have more to do with informatics than biology.

Here is a problem it's a while I have tried to solve with different Linux commands, but I suppose that a script is needed.

I have alist of sequences in FASTA, all with a header XLOC_xxxx. Now, most of them have the same header. I'd like to sort only the first one for each different header (that is, avoid redundancy).
What do youpropose to fix the problem? (Yes, I already have the list of the headers I want to extract).
Thank you in advance!

[GenoMax]

(Yes, I already have the list of the headers I want to extract).
Thank you in advance!

If you already have the list of headers you want to extract then use faSomeRecords utility from Jim Kent's collection: http://hgdownload.soe.ucsc.edu/admin.../faSomeRecords

faSomeRecords - Extract multiple fa records
usage:
   faSomeRecords in.fa listFile out.fa
options:
   -exclude - output sequences not in the list file.

Now for the first part. Are you saying that there are multiple sequences that have identical headers in original source file? Are the sequences themselves identical in that case?

[Henry_C]

Thank you so much for your answer. Well, the initial part of the header is identical, not the last one (in fact, they are different transcripts coming from CuffDiff). But I'll need to keep just one to be able to easily combine the annotation I'll get with their expression profile, in an Excel file.

[GenoMax]

Are these multiple samples on which you ran cufflinks? Did you do cuffmerge first before using the merged GTF file for cuffdiff analysis?

[Henry_C]

Are these multiple samples on which you ran cufflinks? Did you do cuffmerge first before using the merged GTF file for cuffdiff analysis?

Yep, I run cuffmerge first and then cuffdiff using the merged gtf. I have seen people skipping cuffmerge and fusing together all the cuffdiffs together, but then one should use HtSeq Count, probably