Weird alignment results

[bel012]

Hi,

We have very peculiar results that we are looking for ideas on what might cause such results. We have a rat RNAseq experiment with 3 groups and three replicate per group. The RNA, library, and sequencing data quality was highly similar and good in all nine samples. However when aligning reads to RNA genome obtain about 8% of sequence mapping to multiple location in the experimental groups, i.e. 6 samples, and 20% in the three control samples.

We have done some preliminary analysis on results and found the following:

a. While in the two experimental groups about 75% of the sequences map up to 6 different genomic locations and 25% more than 6 locations, in the controls the partition was about fifty-fifty.

b. Next only the multiple aligned reads were counted and while in the two experimental groups 60% of the read were no-feature about 80% no-feature in controls. I.e. this could not explain the over 20% multiple reads in the controls.

Can anyone suggest what might contribute to this difference in mapping to unique location?

Thanks a lot

[dariober]

Hi,

We have very peculiar results that we are looking for ideas on what might cause such results. We have a rat RNAseq experiment with 3 groups and three replicate per group. The RNA, library, and sequencing data quality was highly similar and good in all nine samples. However when aligning reads to RNA genome obtain about 8% of sequence mapping to multiple location in the experimental groups, i.e. 6 samples, and 20% in the three control samples.

Could it be that the differentially expressed genes happen to have repetitive sequences? If one or more genes overexpressed in the controls have repetitive sequences, then you get more multimapping reads in the control libraries. (What do you mean by RNA genome? Did you align to the genome with a splice-aware aligner like tophat? Maybe positing the actual numbers you refer to would help.)