Can GAIIx introduce a bias in coverage A on T?

[vjimenez]

I have a graphic that show an increase in coverge of A over T, and It increses until the end sequence, in all lanes of paired ends runs of 72 pb. Can GAIIx introduce a bias in coverage A on T?

[tonybolger]

I have a graphic that show an increase in coverge of A over T, and It increses until the end sequence, in all lanes of paired ends runs of 72 pb. Can GAIIx introduce a bias in coverage A on T?

Yes, for some reason, our recent datasets seem to favour A over T and G over C, when the quality is a bit crappy. We see a rise in N calls roughly correlating this discrepancy.

An older dataset (of lower quality) of our is different, with C considerably over G, and T slightly over A.

I don't know if this is an effect from the base-caller or if it's from the adapter sequences, but it's there, around 1% difference. How much of a gap are you seeing?

[vjimenez]

The gap between A and T is between 1% to 4% at the end of the sequence. But always in genomic apliaciones of eukaryotes.

[tonybolger]

The gap between A and T is between 1% to 4% at the end of the sequence. But always in genomic apliaciones of eukaryotes.

Probably nothing too serious - how's the quality at that end?

I guess if you're only seeing it in genomic data, it could be related to the specific adapters used.