Difference in count values depends on tool and reference genome version

Question

[shome]

Can counts get different for same fastq files if we used different versions of star and rsem tool for aligning and quantifying.Also if reference genome is different version ?

[dthorbur]

Actually formally tested the importance of different reference genomes generated from different populations using the same samples. Paper here.

It's pretty alarming how different the final results were just using different references. The outcome will be less dramatic for different updates of something as refined as the human reference, but will be large for most other updates.

Answer 1

[ATpoint]

Yes, all factors can play a role. Not sure what more to comment. Reference genome probably the major factor.