Visualization of data obtained from target enrichment approach.

Question

[Artem Kiselev]

Dear all,

I'm looking for tool or pipeline for visualization of NGS data obtained from target enrichment approach.

Our custom panel consists of approximately 100 genes, and I must find a way to visualize the result in an understandable for clinicist form. It seems to me that the most appropriate tool for that is the circos chart, but I have no idea where to start and what tools I should handle our bam/bad/fastq data.

If you can help, here is drawing, as I see it:

Thanks.

[PoGibas]

Artem, if you are familiar with R try circlize (http://cran.r-project.org/web/packages/circlize/), it's R package for circos like figures.

[Artem Kiselev]

Thank you for a concrete reply, now I began to understand that the main problem is not in visualization (there are many tools for that), but in the data preparation process. And the question is: using what methods and in what format should I convert my data so the result will be the information suitable for visualization?

[Alex Reynolds]

The best thing you can do is learn Circos. It is the original and provides the most functionality of all the "circularized" visualization packages. The documentation explains precisely how to format your data and there are numerous tutorial examples provided to demonstrate all of the various visualization subcategories.

Answer 1

[Sean Davis]

I'm not sure that circos is going to give you the resolution that you need. I'd think about using IGV or one of several javascript browsers:

http://www.google.com/search?q=javascript+genome+browser

You can make bed tracks associated with the different classes of variant and bigWig (or tdf for IGV) files for coverage.

If you are really doing this for clinical purposes, you are going to need to spend A LOT of time on reporting strategies. The reporting of clinical results is not a science at all and requires close interaction between the clinicians, the bioinformatics folks, and the lab folks. I cannot stress enough that clinical reporting is non-trivial and requires a different thought process than reporting research results.

[Artem Kiselev]

We really have specialized reporting software, where clinicists are working with NGS data, data are presented in tabular form (with the ability to simultaneously view in IGV).

But we need the overall picture, we need to see what areas are enriched better or worse, which gene has more SNP, where we can still work on our design, etc. For this purpose the circos chart approch and resolution will be enough.

[Sean Davis]

In that case, you'll have to dig into circos (or another implementation) and build what you want, I think.