I am trying to align NGS data to repetitive regions. Straight-forward use of tools such as BWA fails. Any suggestions?
Hi,
If you are working with ChIP-seq data, this article describes how to do it
(http://www.ncbi.nlm.nih.gov/pubmed/21873999)
----------------------------------------------------------------------------------------------------------
Short reads are aligned with BWA such that non-unique reads are reported in your SAM file. After that, non-unique reads are re-allocated to the most probable location based on which of the candidate loci has higher number of unique reads reads aligned to it. You can ask the author for the code.
Alternatively, you can try GNUMAP (http://dna.cs.byu.edu/gnumap/)
Good luck for your work.
What does "fail" mean? It should report the position of "suboptimal" mappings (with sampe or samse -n option). See http://seqanswers.com/forums/showthread.php?t=7129 for details.
Fabio was right to ask what "failed". Seems that what failed was the paremeters I was using. With the right parameters, BWA aligns across repeats very well (at least for the simulated data I am using)
Yusuf: the question is how to align to repeats, not how to avoid the problems caused by repeats. Anyhow, I found that BWA probably gets it right most of the time.
Hi Eitan,
I was wondering what parameters did you use to do the alignment. I am giving the option -n 10 for sampe, but I still find many read pairs that do not align (~95%). Interestingly I also did not find reads with soft clips. Instead when I do BLAT I see that they have hits with last 55 bp of the read aligning with 100% identity.
I had used the commands
bwa aln bwaindex R1.fastq -t 8 > R1.sai
bwa aln bwaindex R2.fastq -t 8 > R2.sai
bwa sampe bwaindex R1.sai R2.sai R1.fastq R2.fastq -n 10 -N 20 | samtools view -bS - | samtools sort - sample_sorted
Thanks
- Badges
- Science method