feature modification + addition

[ramaniak]

Hello,
I am using treeomics and find it is stunningly fast. I have a couple of questions regarding features and execution

1. Is there a default threshold for considering a variant to be present? For example, if a variant in one of the samples has only 2 alt reads over a read depth of 60 reads would it be considered? How would I go about changing this threshold?
2. I notice that some genes are displayed by default along the branches of the tree. Is it possible to change the set of genes displayed? How would I do this?
3. The circos plots that are generated do not display gene names either. Is there a way to do so?
4. This is more of a feature request: is there any plan to work on a version where CNVs can be incorporated?

thanks
Arun

[johannesreiter]

Dear Arun,

Thank you for your email. We just uploaded version 1.7.3 of Treeomics.

1. There is no default frequency or variant read threshold. You can play with the parameters for the Bayesian inference model (defined in settings.py: BI_E = 0.01 (sequencing error rate), BI_C0 = 0.5 (prior mixture parameter of delta function and uniform distribution), MAX_ABSENT_VAF = 0.05 (maximal absent VAF before considering estimated purity). Moreover, the probability if a variant is present or absent will depend on the estimated purity (described in the paper).
2. In the newest version, a gene name is depicted if there is a non-synonymous mutation in a putative driver gene given in the file DRIVER_PATH = '../input/Tokheim_drivers_union.csv' again in settings.py (union of driver lists inferred by 20/20plus, TUSON and MutsigCV in Tokheim et al, PNAS, 2016). You can of course generated your own list of highlighted genes by providing a different file.
3. Gene names are not displayed because each node is a combination of multiple mutations and each mutation contributes some likelihood to each node. One could generate a circos plot with just the most likely mutation pattern for each variant but we haven't done that.
4. There are plans to integrate CNVs in the future but this will not happen before the end of the year.

I hope this clarified your questions.

Best wishes,
Hannes

[ramaniak]

Thanks for your response and I will try out the latest release. Please let me know when you have any updates on CNV incorporation, I would very much love to test this out.

thanks
Arun

[ramaniak]

One more question:
Is there an estimate of clonality checked or reported in any of the outputs? if not, would it be possible to estimate using the existing outputs?

thanks

[johannesreiter]

If you mean with estimated clonality, neoplastic cell content then these estimates are part of the HTML-report that Treeomics produces automatically.

[ramaniak]

Hello, Just curious to see if you have an update on including CNV calls into treeomics?
thanks
Arun

[johannesreiter]

Hi Arun,
sorry, no update regarding the integration of CNV calls yet. Hope to have some news in fall.
Best wishes,
Hannes