What tools to map partial reads on genome?

Question

[biolab]

Hi everyone,

As my title states, could anyone give me some suggestions about partial reads mapping? For example, a reads is normally 100 bp long, however, I would like to find the mapping loci with 20-100 bp range. Actually I tried BLAT(blat -t=dna -q=dna -minIdentity=96 -stepSize=5 -maxGap=0 genome.fa reads.fa output), but it seems to miss some loci.

Your advice will be much appreciated. THANKS a lot!

Answer 1

[Irsan]

What you want is called local alignment and is implemented in for example bowtie2. Depending on your application you will have to optimize the parameters. Sometimes it helps a lot to know the context of your question. If you explain the background of your question you might get better answers

[Devon Ryan]

It should be noted that local alignment is a bowtie2 (rather than bowtie1) option.

[Irsan]

Thanks, I changed it

[biolab]

Hi Irsan, thanks a lot for your comments. Actually I am working on gene splicing in a non-model organism. The genome has tremendous of scaffolds. Of particular interest, genome rearrangement occurs. Would you please inform me the bowtie parameters for local alignment? I will try and compare with BLAT. THANKS!

[Irsan]

You mention your are working on gene splicing, have you done RNA-sequencing or DNA-sequencing?

[biolab]

Yes, we have RNA-seqs of the wt vs mutants. Draft ref genome is known.

[Sam]

Then why don't you just use the RNA Seq alignment tools like STAR or TopHat? They should be able to handle splicing for you

[Irsan]

Yup I was about to say the same thing. Use a dedicated RNAseq mapper like STAR, Tophat2, MapSplice, GSNAP, ...

[biolab]

Thanks, Irsan, these tools are useful. I will try them.

[biolab]

Hi Sam, thanks for your comments. I did not tried STAR. I did try bowtie with default (or I added --very-sensitive options) settings. However, bowtie could not detect some novel srambling rearrangements (I think I need to read bowtie manuals, as Irsan mentioned local alignment). By performing reads partial mapping, these events in mutants could be revealed. It's not normal alternative splicing. The annotation is not complete also. I appreciate anyone's advices. THANKS.

[Sam]

Then maybe GSNAP will suits your need most. They support the read partial mapping as long as you provide them with a GTF file. It can also support the alternative splicing so it does seems like it will fits all your need.

[Irsan]

STAR en Tophat2 also do local alignment en gene annotation guided mapping