Remove identical mapping reads between two indexed samples

[sarvidsson]

bedtools intersect -f 1 -r would give you an idea but won't handle paired reads, so you will have high rates of falsely accused carry-over fragments. You would need to filter the input and run it several times, each time exporting the individual IDs of your sequences, and compare those lists in the end. Possible, but tedious.

Another way to get a rough idea but avoid alignment would be to merge the read pairs (if the fragments you've sequenced are shorter than <580 bp this works well), awk for the sequence lines, sort and get the count of unique sequences per sample. Then write a small script to compare the unique sequences between samples.

In any case, you will falsely accuse some reads at high coverages. If you are really interested in getting quantitative values for carry-over contamination, you should spike in different synthetic DNA in each sample...

[GenoMax]

See this: http://seqanswers.com/forums/showthread.php?t=39270

Post #3.

I think you should be able to provide multiple files as input (separate them by a comma). Brian can confirm.

[felvis56]

I have 24 indexed samples from a single Miseq run and I want to remove identical reads between different samples i.e. cross-talk or potential lab contamination/carry-over.

I was trying bedtools intersect -f 1 -r to try and report only unique sequences from bed file but I'm not sure this is the right tool for the job.

Thanks in advance!