Hi, I have RNA-seq data (from ribodepleted samples) and i'm trying to map my reads to features using either htseq or featureCounts. The input for both is a bam file (reads were aligned using STAR) and a GTF file from ENSEMBLE. Both assign reads to exons by default.
FeatureCounts reported that ~33% of my reads were assigned to features, however when I'm trying to check how many reads were assigned using htseq, I saw that only 2% of my reads were assigned (There is no report in htseq so I summerized the second column in the results file, without the last 5 rows that contain the "no-feature" and "ambiguous" etc).
What is causing this difference between the two algorithms? Thank you very much!
1 answer
•
written
by
Arup Ghosh
Log in to answer this question.
-
FeatureCounts: Low assigned reads, can I proceed? Do I need to do another pre-processing step?
written by beginner123 •
-
FeatureCounts output and downstream analysis
written by kb_93 •
-
R featureCounts very low assigned read rates
written by 1111111 •
-
Using StringTie output GTF with featureCounts to assign reads- low assigned percentage
written by sg197 •
-
Why I have too many reads non uniquely aligned?
written by zizigolu
-
Is Gene's counts done correctly?
written by Nick •
-
featureCounts: 0% successfully assigned fragments on multiple PE .BAM files
written by imesi •
-
FeatureCounts on HG38
written by elb
-
Warning when using gff format in featureCounts - miRNAs
written by MeiNB •
-
Extremely low assigned reads in featureCounts
written by xqyjxau •
The differences should be very minor between the two tools. If you post the exact commands used, it will be easier to see what causes the differences.
Is it possible you're using the wrong stranded argument in htseq? The vast majority of stranded libraries (e.g. using the Illumina kit) should use --stranded=reverse NOT --stranded=yes
Thank you very much, this solved the problem.