Sample sequencing performence depending on index?

[crisime]

Hi,
I have some Illumina exome data from different runs on NextSeq. I see quite some difference in the number of produced reads per sample, which seems to correlate with the used index.
It occured using different batches of chemicals and I ruled out positional effects in wet lab.

Has anyone run into this problem before?

[GenoMax]

Were the libraries made at different time points or at the same time? By the same person using the same protocol?

[crisime]

Hi GenoMax,
different time point, different persons, same protocol.

[GenoMax]

Then you can't really say for sure that the index is the culprit (perhaps you put that reason down since you can see it is the results). It could be related to library made by a different person and/or other factors.

[crisime]

Sry. I should have answerd more precise.
For each single run all libraries were prepared just before the run was started by just one person. The person producing the library was different for different runs, but not for all.
So there is no connection between MID and Person or MID and timepoint which could explain the effect.

From your answers I conclude that you have never seen such an effect

[nucacidhunter]

If different captures with the same index tend to have lower reads than others it could be due to multiple G residues in index sequence which would affect quality of index read and result in read loss during demultiplexing.

[GenoMax]

If @nucacidhunter is right then these libraries if run on a 4-color sequencer will look equivalent. Something you will have to test.