RNASeq lane effect

Question

[roncalli]

Hi All,

I am looking for an answer..Maybe easy for some of you! I have 2 RnaSeq datasets processed in the same way (same individual, same RNA extraction, cDNA prep and sequencing platform) the only difference is that I ran the 2 datasets on 2 separate lanes. The samples have been multiplexed and have a similar sequencing depth (# of reads). What I am trying to do is to run a gene expression analysis on them. Basically I did map both datasets against a reference transcriptome (previously generated) and I am using the raw count as EdgeR data to run the statistics. My question is...How do I account of the lane difference? Is there a way of quantifying it? Or...does the TMM normalization includes the possible lane error?

Thanks for the help!

Vittoria

[Satyajeet Khare]

Any particular reason you expect lane differences? I have seen technical replicates on neighbouring lanes and they look identical.

[GenoMax]

If the same library was run on two lanes then there should be no difference in the data (read numbers may be more or less if the loading concentration was different).

[stolarek.ir]

there might be something off if you multiplex hardly, and the indexes used within the lane do not co-op with each other well

Answer 1

[zzqr]

If there are two different samples ran on two lanes, the seq data could not tell the difference from lane effect. Have technical replicates might be a good idea to account of the lane difference. The TMM normalization estimates scale factors between samples. It could not distinguish the effect bewteen sample diffence and lane difference.