finding deletions of length 20 - 50bp from 100bp reads

[hrseq]

Hi

What would be the suggested approach to find deletions in the range of 20 - 50 bp in human data? The reads are 100bp illumina reads.

Any suggestions?

regards
hubet

[HESmith]

We have used a split-read strategy (described here) to find indels in this size range. Essentially, single reads are treated as pseudo-paired end reads, which provides single-base resolution for the size of the intervening sequence.

[shi]

Hi

What would be the suggested approach to find deletions in the range of 20 - 50 bp in human data? The reads are 100bp illumina reads.

Any suggestions?

regards
hubet

You may try Subread (set '-I' to be 50). When the specified indel length is greater than 16, it performs local assembly to find long indels.

[gringer]

You may also be able to shoehorn Tophat into doing this as well, because it also maps read subsequences.

[alexdobin]

Any RNA-seq aligner that can detect novel non-canonical splices can be used to detect deletions. I would recommend STAR, of course . The problem with using TopHat for this puprose is that it only detects "canonical" and "semi-canonical" intron motifs (GT/AG, GC/AG, AT/AC).

[jp.]

Hi
I got something very basic and important before alignment. I hope that someone would please let me know this.

I have got raw sequence data (seq1.fastq seq2.fastq) from paired end Illumina2000 company and I used these Paired-End files directly to align with hg19 using BWA and Tophat(bowtie). I checked using grep and found that there enough N. I am wondering if this will impact on my allinment and subsequently on mutation/tophat calls.... OR a trimming is required before aligning these to remove N and then use in BWA ? Can BWA alone handle N and remove them, however, these will not aligned with hg19 so will automatically be removed... ??
Expecting kind reply
Thank you in advance
jp.

Any RNA-seq aligner that can detect novel non-canonical splices can be used to detect deletions. I would recommend STAR, of course . The problem with using TopHat for this puprose is that it only detects "canonical" and "semi-canonical" intron motifs (GT/AG, GC/AG, AT/AC).

[gringer]

Can BWA alone handle N and remove them, however, these will not aligned with hg19 so will automatically be removed... ?

This is unrelated to this thread topic, and also unrelated to the post that you quoted. Please post your question in a new thread.