Error message for too few reads
alvinwt opened this issue · comments
Hi, I have been getting
Traceback (most recent call last):
File "/data/rozencompute2/a0073895/optitest/OptiType/OptiTypePipeline.py", line 304, in <module>
"in your config file (currently %.3f), because you may need to resort to using unpaired reads.") % unpaired_weight
TypeError: not enough arguments for format string
even when I have changed the unpaired weights in the config.ini file to 1. I am wondering if this is a error message when there are few reads in the input fastq or if something else is happening.
It should still run through, it was just an unescaped percentage sign in a string substitution one line above, fixed now. BTW, why do you have so many unpairable reads? Could you run this with deletebam=false in the config file and e-mail me the intermediate bam files and the coverage plot? I've never come across a sample like this (which is the reason that unescaped % went unnoticed for so long :))
And forgot to answer your question: you get this warning if your HLA reads' individual ends can't be paired up consistently (i.e. both ends mapping to at least one common allele) for more than 10% of them. I could only see this happen for very long insert sizes where one end always falls in the exon2-exon3 reference window but the other is always left out.