Reason for low quality of illumina reads

[nvteja]

HI All,

Can anyone let me know why is it the case that the read quality is generally lower towards 3' end for solexa ?

Thanks,
Teja.

[Torst]

Can anyone let me know why is it the case that the read quality is generally lower towards 3' end for solexa ?

Put simply, the reads are synthesised base by base from the 5' end. Each step is error prone, hence errors accumulate. Each step makes the molecule longer and wobblier, making it hard to do image processing.

[SillyPoint]

Another phenomenon at work here is what Illumina calls 'phasing'. Remember that each cluster contains ~1000 (supposedly) identical molecules, all of which are undergoing reverse-strand synthesis.

In a perfect world, the chemistry works perfectly, and at each cycle each molecule has one base+dye+blocker added, and has dye+blocker removed after imaging. In the real world, a small fraction of the molecules doesn't come out right: they miss a cycle, or get an extra (unblocked) base added.

The cumulative effect of phasing over many cycles means that eventually the cluster has so many molecules out of sync that there is no clear base call possible, and miscalls start to happen. Quality scores, which are derived from the ratios of the intensities start to degrade. Eventually, you have to stop.

--SP