a lot of multiple alignment reads after rRNA depletion

[chenyudr]

I met the situation of high % reads with multiple alignment sites. Our procedure is removing rRNA by Ribozero kit from ~1 microgram RNA isolated from rat liver tissue followed by standard RNA seq library construction. In some samples, we found a lot of reads, ~40-50%, has high similarity with rRNA and other ~20% reads with multiple alignment sites, of them most belong to 7sRNA.
We also checked bioanalyzer map of RNA after rRNA removal and found a ~140bp peak with with high FU value ~200. However we didn't check peaks around 18s and 28s positions in the map.
We speculated that the tissue has kind of specific features that Ribozero kit can't remove 7s RNA or SnRNA thus result in this multi-alignment situation.

Has this happened in your experiment? Could you give me some suggestion?

Many thanks,

Yuk

[dpryan]

The ribozero kits are only ever "OK". We pretty commonly still have up to 20% rRNA (rarely higher, though it happens on occasion) and this always causes the multimapping issues you're seeing.

[chenyudr]

Yes, but we didn't find similar situation in samples from other tissues.

[dpryan]

Then perhaps the kit is indeed missing something. Some tissues do have different rRNA composition (e.g., sperm), so it can happen.