replicate1 __no_feature 39795601 __ambiguous 73269 __too_low_aQual 2553578 __not_aligned 822830 __alignment_not_unique 6895933 Is this result normal? I used HIAST to do alignment. Because my sequencing data is Sanger/Illumina1.9, so I only set the quality value as Phred 33 and run other parameters as default. My alignment rate is about 90%. However, the HTseq result is quite unexpected. I have no clue where went wrong.
• 0 views
•
link
written
by
sophialovechan •
Did you map the reads to the genome or transcriptome?
it was mapped to the genome using tophat2
So you should use -s yes. The differences you show are expected even with random reads
That's even not something he tested and will for sure depend on the protocol. You can have stranded or unstranded RNA seq.
This are the results from
-s yes:It looks like it's stranded but as written below, you have to make sure which protocol was used.
A huge number now ends up in no_feature which argues against -s yes. My guess is non-stranded, in which it makes sense that there are more __ambiguous reads (which can not be assigned since htseq-count has no strand information available). ambiguous means that the read can be from either geneA or geneB, without strand information impossible to say in the case of antisense transcripts.
I missed a digit :) In that case it's probably reverse since most of the reads do map in reverse mode