customized coverage report give back "0/0 " transcipt for a gene present in the panel

[4WGH]

I tried to make a custumized coverage report for a gene (NR2E2, HGNC ID 7974)

I get a 0/0 transcripts ,see image

ad the gene is present in the panel, and I can see in IGV it is covered I do not understand the zero at the denominator:
in 0/0 (= x/N) I can understand if x=0 but N can not be 0.

I did test with other genes and I get a proper result.
I tested this specific gene in another patient and I get 0/0.

Can you help me? a bug or I'm missing something?

[4WGH]

i also see this message in the main page

[northwestwitch]

The gene might be missing from chanjo database for some reason? We'll check. Thanks for reporting!

[northwestwitch]

Mm I can't connect to the chanjo database any more. Something must have changes in the settings..

[dnil]

Hmm, old chanjo. This is not so hot:


I guess one has to pay careful attention to that "no matching transcripts found" flash, huh.

[northwestwitch]

Password was changed some months ago. Anyway, no transcripts from that gene in the database:

[northwestwitch]

I'm moving this issue to the chanjo repository. How easy is to update the transcripts in there @dnil ?

[dnil]

Not exactly easy, hence partially chanjo2. But nothing is impossible of course.

[northwestwitch]

A dirty way would be adding it manually..

[dnil]

Yes, we would just need to make sure we are not causing some issues in other areas, in particular for old samples that do not yet have entries for these new transcripts?

[northwestwitch]

Weird that it's not present. I wonder why!

[4WGH]

Some follow up questions:

1. This means the coverage report for the panel that contains this gene is not correct (in big panel the effect is not big but it could be in small ones).
2. how do we know if there are genes in my panel of interest that are not in the Chanjo repository?
3. Why the gene is not in the repository?
4. I suppose variant calling in this gene is not affected ?

[northwestwitch]

I'll try to answer point to point:

1. I think this is rare case. The coverage report will contain the data from all the other genes and the effect would be minimal for larger panels but of course if you have a panel of 2 genes the report will be based on the other gene that is not missing. - Keep in mind that this means that the gene was not found in the software to create the report, not that it's not covered . And in fact you see the coverage on igv -
2. You got that warning in scout "no matching transcripts found", that's a good indication.
3. I don't know. I asked first! 😆
4. Nope, all good by that point of view

[dnil]

Sry, listening to a meeting, but my 5 cents quickly.

1. Exactly, panel coverage is for genes found. This is known for eg genes missing from hg19.
2. You would see them as 0 transcripts and get a flash warning from chanjo... :/
3. good question, we are investigating. It is a gene family of sorts, so there may be something that caused it not to be in the gene set some years ago when chanjo was initialised
4. No, variant calling is not affected. It's all good!

We have a known limitation in updating the chanjo gene and transcript sets; the difficulty of this was part of what made us develop chanjo2, which will likely launch with the new nf-core/raredisease pipe - later this year?

[4WGH]

thanks! one more question following point 2
If I hava a panel (ex Retina distrophy) and want to check if any other gene is not in Chanjo repository, how do I do?
I tested to select "show genes" in the custom features leaving in all panel genes and this is what I get

as the limitation is to 50 the NR2E3 is not even listed.

anyway if we move to Chanjo2 I can survive with this question unsolved.

[northwestwitch]

thanks! one more question following point 2 If I hava a panel (ex Retina distrophy) and want to check if any other gene is not in Chanjo repository, how do I do? I tested to select "show genes" in the custom features leaving in all panel genes and this is what I get

as the limitation is to 50 the NR2E3 is not even listed.

Don't confuse the genes that are not covered above threshold (those max 50 genes in the cell) with those that can't be evaluated because not present in chanjo. You don't know if NR2E3 would be part of those not fully covered above threshold because NR2E3 was not evaluated at all. Could also be that NR2E3 is super-covered, but you don't know because it's missing from the database.

anyway if we move to Chanjo2 I can survive with this question unsolved.

We are working on it and it will be available in production relatively soon I think! 🤞🏻

[dnil]

It also would really not hurt if chanjo would report in detail genes in the query which do not have transcripts in its database, but having chanjo2 is higher prio by far, for this reason and many others.