Hi all,

I have paried end RNA seq data. After doing alignment with STAR, I used featureCounts (in R) to count genes.

When I used the annotation file download from Illumina iGenomes UCSC mm10.gtf, I got very low (19309701 (22.7%) percentage of successfully assigned fragments (see figure in the left part).

Then I used the default annotation file from featureCounts. The percentage of successfully assigned fragments got higher from 19309701 (22.7%) to 36651127 (43.2%). 

I also downloaded gtf file from Ensembl the successfully assigned fragments were still lower than the default.

Can someone help me better understand why there are so much difference in the percentage of successfully assigned fragments when using the mm10 gtf files from different resources?

Is there a good platform to download the "good" gtf annotation files?

Any advice is appreciated.