[maker-devel] isoforms

[Carson Holt]

Sorry for the slow reply I’ve been away this last week. Thee is no parameter for isoform strength per se.  The ability to call isoforms is strictly determined by the strength of evidence you have.  Basically The gene predictors are iteratively ran with a single piece of EST evidence being primary and the remaining evidence being secondary, and then the gene predictor can make any changes it deems appropriate.  Most of the time the exact same model comes back, but if a particular piece of evidence suggests a novel splice site then a new model can be produced based of of that hint.

However if your EST/mRNA-seq evidence has a lot of noise or contamination, then you may be feeding in a lot of bad hints. These may get ignored since they would generate unworkable ORFs, but not always.  There is unfortunately no good way to automatically distinguish a good hint from a bad hint. However if you run MAKER’s results through EVM (Evidence Modeler) you can manually assign weights you deem appropriate to each evidence source. EVM can then modify models based on these weights.

—Carson


> On Sep 21, 2015, at 6:48 PM, Parul Kudtarkar <parulk at caltech.edu> wrote:
> 
> Hi,
> 
> Is there any parameter to be used while running MAKER2 pipeline to filter
> out weak isoforms?
> 
> Thanks,
> Parul
> --
> Scientific Programmer
> Center for Computational Regulatory Genomics
> Beckman Institute,
> Biology and Biological Engineering
> California Institute of Technology
> http://www.echinobase.org/
> 
> 
> _______________________________________________
> maker-devel mailing list
> maker-devel at box290.bluehost.com
> http://box290.bluehost.com/mailman/listinfo/maker-devel_yandell-lab.org