Very Low Coverage NGS Copy Number analysis

[ENT Fellow]

Hi all,

I'm doing very low coverage sequencing using the Illumina HiSeq running 40/lane looking at DNA copy number in FFPE head and neck cancer samples. I've been using CNAnorm to analyse the data, create karyograms etc. as this seemed to be specifically designed with low coverage NGS in mind and has ben published in cancer studies.

With so many software packages available does anyone have any experience of other packages with specifically this low coverage NGS in FFPE DNA CNV? I'm a bit concerned at having to defend the choice in my viva.

many thanks

[ni16]

Hi ENT Fellow,
What if alongside some low coverage area, I want to know copy number of higher coverage area, would you expect CNAnorm to equally good, or you see some issues with the tool?

[Richard Finney]

I've used bic-seq2 with low coverage with good results.
Note that you need to adjust lambda parameter for low coverage samples.
It's a bit of hassle to setup and get going.

[Brian Bushnell]

I originally developed BBMap for a lab with very low-quality and low-coverage data. Not specifically for karyotyping, but for identifying mutations in low-coverage data, by rigorously optimizing the true/false positive/negative rates in low-coverage data. As such, it tries very hard to make the most of all input data.

BBMap does not do karyotyping, directly, though it can directly output coverage information. But it is extremely accurate in the presence of rampant mutations - including 100kbp+ deletions, which might erase an entire suppressor gene - and as such, is ideal for cancer experiments. The ability to directly output per-base coverage might be helpful as well.

[zajic]

Very Low Coverage NGS Copy Number analysis

Hi, I used 2 R libraries: QDNAseq and ReadDepth. Both seem to have similar results.