Question regarding reported PFAM domains and coordinates in STAR-Fusion --fusioninspector validate results

[Martin]

Hi all,

Thanks for creating such a useful set of tools to identify gene fusions.  I want  to ask you a few questions regarding the coordinates reported on the "star-fusion.fusion_predictions.abridged.coding_effect.tsv"  file generated after STAR-Fusion runs with `--FusionInspector validate --examine_coding_effect --denovo_reconstruct`   parameters using the CTAT library (GRCh38_gencode_v37_CTAT_lib_Mar012021.plug-n-play).  The reason for this is because I'm trying to write some code to plot the retained PFAM domains by inframe fusions over the CDS while highlighting the breakpoint, which gene (either left or right) the PFAM domains came from.

I'll be using the example fusion on the githubTutorial wiki, ARFGEF2--SULF2 :

• CDS_LEFT_RANGE (eg. 1-121) and CDS_RIGHT_RANGE (eg.176-2610) :   I infer that the these coordinates shown here are based on the position on the CDS sequence reported on the FUSION_CDS column. Is this the case or are these based on the coordinates on the original CDS for the transcript annotated on the respective CDS_LEFT_ID & CDS_RIGHT_ID?
  • I'm currently using the informationt that CDS_LEFT_RANGE provides to plot as the position of the fused breakpoint on the FUSION_CDS (i.e. 121)
    
• PFAM_LEFT and PFAM_RIGHT:  Regarding the coordinates reported here I have two questions: 
  • Are the PFAM HMM results on PFAM columns result against the fused protein or the original individual proteins produced by each transcript ?
    
  • Are the coordinates reported for the matching PFAM domains related to positions on the FUSION_CDS column or the protein sequence on the FUSION_TRANSL column or the original transcript CDS_LEFT_ID  & CDS_RIGHT_ID? 
    • The reason for these question is because I have noticed that the coordinates for the PFAM domains if they were protein some of them would be outside of the location contributed by that gene to the FUSED_TRANSL protein.
      • • Eg. for  ARFGEF2--SULF2, CDS_LEFT_RANGE=1-121, which would be only 40 aminoacids, and  PFAM_LEFT=DCB-PARTIAL|23-121~|1.8e-41^DUF1981|69-95|0.37^HEAT|73-95|0.013a

Many thanks for all you time and help.

Hope you have a good day.

Best regards,

Martin

[Brian Haas]

Hi Martin,

The Pham hits are precomputed based on the reference annotations and the coordinates should be based on the original transcript structures and sequences as well. It only takes the breakpoint into account and tries to find in-frame fusion proteins based on that breakpoint and reference structures.

If needed I can dig into the code to pull out more specifics.  I expect it’s part of the ctat genome library construction steps.

All the best

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

 

--
You received this message because you are subscribed to the Google Groups "Trinity_CTAT_users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trinity_ctat_us...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/trinity_ctat_users/936bf0fa-3449-4971-9f3c-0ceb8659555cn%40googlegroups.com.