How to resolve differences with the web installation

[Steven Handerson]

I'm seeing a slight difference in the ansible mutalyzer I built vs. the web version --

specifically,using the position converter with hg19 gets different results.

You can treat this as a newbie question as to how to get exactly the same setup as the web site --

maybe a request to post all maintenance operations on the mutalyzer so we can duplicate them

in order to get the same results.  First is the web results, then our results.

Maybe a walkthrough as to how to resolve this particular issue might let me generalize.

Query:

chr11:g.66328095T>C

Chromosomal variant

NC_000011.9:g.66328095T>C

Found transcripts in variant region

ACTN3     :	NR_047663.1:n.1826T>C
		NM_001104.1:c.1729T>C
CTSF      :	NM_003793.3:c.*3309A>G

Chromosomal variant

NC_000011.9:g.66328095T>C

Found transcripts in variant region

ACTN3     :	NR_047663.1:n.1826T>C
CTSF      :	NM_003793.3:c.*3309A>G

[J.F.J. Laros]

Dear Steven,

This is probably because your database content is different from ours.
Mutalyzer is updated on a regular basis with new transcripts from the NCBI, but
it also keeps the transcripts the NCBI no longer provides.

I hope this helps.


With kind regards,
Jeroen.

> --
> You received this message because you are subscribed to the Google Groups "Mutalyzer" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to mutalyzer+...@googlegroups.com.
> To post to this group, send email to muta...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/mutalyzer/e44fb02c-fd50-472f-b13b-06886309a33e%40googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.


--
Jeroen F.J. Laros - J.F.J...@lumc.nl

[Martijn Vermaat]

Hi Steven,

Jeroen is correct, but I would like to clarify that the NCBI is not at fault here. They don't change or remove published transcript mappings.

The gist you probably found [1] lists the most recent MapView files available for each assembly. For hg19, that's annotation release 105 (based on GRCh37.p13). This apparently does not contain NM_001104.1, but some older MapView file does.

As Jeroen said, the database on our server is the result of accumulation over time. In addition to importing a new MapView release every now and then, we also sometimes import a transcript from the UCSC Genome Browser database if it is not available in the MapView file. And before using the MapView files, we used to get everything from the UCSC for a few years. Unfortunately, this is a manual process which was not documented or tracked, so there's not really a way to get to the exact same database as ours. I'm planning to tackle (a part of) this by managing these imports with Ansible [2].

So I would say the differences between your installation and ours should mainly be historical. For all latest transcript versions, there shouldn't be a difference. If you would really like to have the exact same database I could provide you with a snapshot as a raw database dump. It might be a good idea to provide such a dump as a service at some point.

cheers,
Martijn

[1] https://gist.github.com/martijnvermaat/ce84945d05b4e42d3584
[2] https://github.com/mutalyzer/ansible-role-mutalyzer/issues/6

> To unsubscribe from this group and stop receiving emails from it, send an email to mutalyzer+unsubscribe@googlegroups.com.