Inquiry about '--examine_coding_effect' parameter

[小林賢子]

Dear author,  

I would like to ask a question.

STAR-Fusion outputs CDS_LEFT_ID and CDS_RIGHT_ID as the results of '--examine_coding_effect' parameter. How does STAR-Fusion automatically choose each transcript? (I mean, there are a lot of transcripts for each genes. What criteria are applied for choosing transcripts?)

Tetsuko Kobayashi

[Brian Haas]

Hi,

It just prioritizes the candidate fusion isoform that produces the longest in-frame coding sequence given the reference annotation structures and the tentative breakpoint.  It doesn't do anything fancy or more intelligent here at the moment.

--
You received this message because you are subscribed to the Google Groups "STAR-Fusion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to star-fusion...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/star-fusion/17e50707-a740-4eab-8fa0-762eedc5ef4cn%40googlegroups.com.

--

--
Brian J. Haas
The Broad Institute
http://broadinstitute.org/~bhaas

 

[小林賢子]

Thanks a lot.

2020年8月26日水曜日 22:08:23 UTC+9 Brian Haas:

[Mario N]

Hello,

About this option, I have trouble to understand the "est_J" and the "est_S" column. It seems to be related to the JunctionReadCount and SpanningFragCount. Can you explain it to me ?

Best,

Mario

[Brian Haas]

hi,

These are adjusted counts taking into account multimappings.  For example, if there's a spanning frag that's assigned to multiple different isoforms with different breakpoints, the est-val involves a fractional assignment of that read to the isoform based on likelihood given all the other read mappings and proposed expression (FFPM) value.

This is a new addition in recent releases.

best,

~b

To view this discussion on the web visit https://groups.google.com/d/msgid/star-fusion/13827f58-7700-404b-8559-786c8180e3e5n%40googlegroups.com.

[Mario N]

I seem to catch the idea but i'm still a bit confused.

From the data I have, the junctions reads are always equal to est_J. Does it mean that I have a single breakpoint on these fusions ?

On the other hand, the spanning reads are most of the time different of the est_S (eg :  210 / 196.60 or 216 /16.16).

So in the first case I have a major isoform and in the another one a lot a differents isoforms ?

Thanks again for the help,

Best,

Mario

[Brian Haas]

The spanning reads are reassigned to the isoform that's most likely the correct one, but they're compatible as evidence with the other one.

Think of it as having the 'raw counts' and then the 'estimated counts' as you would in isoform expression quantification data, in case that helps.

To view this discussion on the web visit https://groups.google.com/d/msgid/star-fusion/93875050-1415-46a2-acd6-f143fe420d2an%40googlegroups.com.

[Mario N]

Oh okay, It's a bit clearer .

Thanks for the enlightenment.

Best,

Mario