High no feature counting when using strand specific data

Question

[hamaor]

As i'm analyzing strand-specific RNA-seq data, prepared with Epicenter kit, I used HT-Seq count with the option: -s yes Some of my samples had only ~50% counted reads (the rest is had no feature) while others over 85%. i tried to count again, this time -s no, than all of my samples had over 85% counted.

Consulting with the library preparation staff, we checked for DNA presence but couldn't found any abnormal levels.

any idea how to move on or what might be the cause ?

[WouterDeCoster]

I have no experience with the epicentre kits, but chances are you need -s reverse (if this kit uses the dUTP method for strand specificity)

Answer 1

[hamaor]

I tried it. counting percentage drops dramatically. though samples has high percentage of "no feature" when strand= yes presents more counted reads than others when strand = reverse.