Telomere filtering of MiSeq PE data

[winsettz]

Working on de novo assembly, but this particular problem is more about pre-filtering telomeric reads (the sequence of the repeat is known).

Any robust workflows to filter 150 nt MiSeq reads for telomeric reads? Or more abstractly, for filtering reads with a particular known kmer of varying repeat length, such as (NNNNN)X? Thanks.

[tonybolger]

Working on de novo assembly, but this particular problem is more about pre-filtering telomeric reads (the sequence of the repeat is known).

Any robust workflows to filter 150 nt MiSeq reads for telomeric reads? Or more abstractly, for filtering reads with a particular known kmer of varying repeat length, such as (NNNNN)X? Thanks.

Trimmomatic might be able to do this. Can you send me some example reads and what sequence you'd like to remove?

[winsettz]

My starting assumption (born out by the literature) is that telomeric sequence is a repeat of "TTAGGG".

The proposed workflow is to pass each fastq file through fastx_clipper

fastx_clipper -Q33 -A TTAGGG -i input.fastq -o output.fastq

I think fastx_clipper is actually removing reads with telomeric reads, and to use it downstream in velvet as paired reads I need to ensure that both pairs are present.

I figured I would borrow khmer's pair-sorting functionality, but it works on interleaved files, so interleave then sort.

python ~/khmer/sandbox/interleave.py read1.fastq read2.fastq > interleaved.fastq

python ~/khmer/sandbox/strip-and-split-for-assembly.py interleaved.fastq