Can we merge 2 csfasta files ?

[tdm]

Hello,
I would like to merge 2 csfasta reads files:
We have performed a poor sequencing.
We re-did the sequencing.
Now we want to run the mapping software on this 2 read sets together in order to increase our coverage, and so the quality of the analysis.

Do you think that this a way to merge the 2 files ?

I have had a look into csfasta and qual files, and i can imagine some tricky think with fasta titles.

Regards,
Thibault

[nilshomer]

Map each separately, then merge the SAM files. The algorithms map each read independently, so you only need to merge the results before variant detection/calling.

[tdm]

Thx for your answer
there won't be any quality or score values bias ?

[nilshomer]

Nope._____

[tdm]

perfect
I will try that next weekend.
Thx for your help !

[westerman]

I'll agree with Nils in that merging the SAM files is the best approach. However this option is only available in the most recent versions of the ABI software.

If you happen to be stuck with older software then you can also combine csfasta and qual files together if you rename each of the reads in them to something unique. This requires a bit of coding but is not hard since the read names are not used by the ABI software. Just keep the overall format looking the same.

[tdm]

Thx Westerman,
This was my 1st thought.
Merging sam files is faster, but merging qual and csfasta files makes more sense to me.
So if i prefix fasta names by 1_ for the 1st run, and 2_ for second run, it will be fine ?

Thibault

[westerman]

The header portion looks like:

>aaaa_bbbb_cccc_F3

Or whatever primer you are using. I just make new header iterating each of 'aaaa', 'bbbb', 'cccc' from 1 to 9999. E.g.,

>1_1_1_F3
>1_1_2_F3
>1_1_3_F3
...
>1_2_1_F3
etc.

Should be a snap to code. The major hassle is keeping the csfasta and qual files in sync.

[tdm]

Should be a snap to code. The major hassle is keeping the csfasta and qual files in sync.

That is why i would just prefix by 1 or 2 the headers, depending on the run (1st or 2nd). So i will be sure it is ok.

[westerman]

Yes, you might be able to do the prefixing. That would be even simpler.

sed --in-place -e 's/^>/>1/' file1

Repeat 3 times for qual and file2.

I am unsure what the limits are on the headers. But it is worth trying.