Hi everyone! I was using fastq-multx to demultiplex and met some problems.In the experiment the two samples were mixed in one sequence run, so I got R1.fastq and R2.fastq with mixed samples, with index and barcode designed as:
sample 1: forward barcode :TATAGCCT , reverse barcode CGAGTAAT sample 2: forward: TATAGCCT, reverse TCTCCGGA
First I use 8bp barcode to demultiplex, but the demultiplex result is bad. 99.99% count were unmatched....
and then I use the full length of index (33 nt, with 8 nt barcode) to demultiplex, and only 8%-10% reads can be demultiplex. ~90% reads were unmatched.
I checked the unmatched reads file, found some reads lack of barcode, some reads contain both forward and reverse barcode, some reads seems to have wrong barcode (e.g. forward barcode in R2 file).
thanks for your recommending!!! I then checked at the unmatched reads and found some R2.fastq reads and some R1.fastq reads contain both reverse and forward barcode. that seems to be why these reads were unmatched. Would this be caused by problems of sequencing? Do you think if I try demultiplexbyname.sh can improve the proformance?
There can be only one version of the index sequence present in the output. If you expect
TATAGCCT+CGAGTAAT
combination then you should use that when you demultiplex the data.If you experienced problems with sequencing that is a different issue. No tool is not going to help with those.