I ran picard's CollectGcBiasMetrics tool for my WES sample, with the intention of assessing the performance of the Exome capture panel in terms of the uniformity of read coverage at various %GC windows. I obtained the gc bias plot attached below. enter image description here


I am a bit puzzled by the normalised coverage values in the windows containing higher GC % and need help with interpretation - in particular, whether it indicates that something has gone awry with the sample, or if it is merely the way the normalised coverage is computed in picard. Why do the normalised coverage values seem to peak at around 50-55 GC%?

Digging into the gc_bias_metrics.txt file: enter image description here

The #windows, #read starts and normalised coverage values are highlighted in the red, blue and yellow boxes respectively. Here, I notice that the # reads / # windows ratio is high for the tail end GC content, which probably inflates the normalised coverage value.

Is this something unexpected for a WES sample, i.e. should I be concerned?

Follow-up question

If the coverage values provided by picard isn't that useful, are there other useful tools for me to calculate the uniformity of read coverage across various GC%? I am looking at plots like Figure 3 in the twist biosciences data page.

  • $\begingroup$ Often coverage is higher in the relatively high GC range due to some artefactual biases from things like PCR, though the exome kits are getting better these days (and can be PCR-free): nature.com/articles/s41598-020-59026-y $\endgroup$ 2 days ago


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Browse other questions tagged or ask your own question.