Understanding paired-end SOLiD

[AmitL]

Hello everyone,
I am working with paired-end SOLiD data, and I am not sure how to handle it.

1) What exactly is the difference between the F3 and F5 reads? Is the F5 reversed?

2) I'm using BWA for the analysis. How do I use it with paired-end? In the documentation it says 'file 1' and 'file 2', but I don't know which is which.
I have tried to run it both ways and in both times I got very few variants - 3000, 5000 while expecting 50000.
I have a color-space reference index, and I used the -c parameter in the alignment. Am I missing something?

3) Does the csfasta conversion work differently with paired-end data?

Thanks a lot,
Amit

=]

[gringer]

1) What exactly is the difference between the F3 and F5 reads? Is the F5 reversed?

They are strand-specific reads. For the same RNA sequence, they will map in different directions to each other.

I think the 5' read is reverse and done first, so if you're doing an analysis taking reads in the usual order you need to specify RF orientation in the mapping program (I'm not very sure on this point).

3) Does the csfasta conversion work differently with paired-end data?

Nope. You shouldn't do this conversion with either. Map in colour-space and let the mapping tool give you the equivalent base-space sequence.

[AmitL]

Thank you very much!

Do you run such analyses with BWA?

[gringer]

No, I haven't [yet] bothered with looking at another program. While I might get a higher proportion of mapped reads, I've got enough problems working with error in the reads I have from Bowtie.

[AmitL]

But... Bowtie does not handle color-space... or does it?

[gringer]

Bowtie can do color-space, Bowtie2 cannot (and is not likely to in the future).

[AmitL]

~bump~
Does anyone work with BWA on SOLiD paired-end data?

[nilshomer]

BWA does not support paired end reads with SOLiD.

[AmitL]

I know the new version does not work with color-space.
Are you saying that it's true for the older versions too?

[nilshomer]

The new version does not work with color space. The older version works with mate pairs and color space, but not paired end and color space.

[AmitL]

How do you know it's meant for mate-pairs?

On the documentation it says:

Generate alignments in the SAM format given paired-end reads.

[nilshomer]

How do you know it's meant for mate-pairs?

On the documentation it says:

Check out the source code "bwape.c" and search for SOLID. It uses the F3/R3 "pairing", which is mate pair.

[AmitL]

Alright. Thanks a lot!

[yksikaksi]

Have you consider to map your csfasta reads with SHRiMP2?

[AmitL]

Yes.
Shrimp2 collapses because of non-existing input problems. I have checked the input manually and it was ok. It works with all other tools.

I preferred not to deal with such a buggy tool.

But thanks!