Retrain using braker.gtf + new proteins

[sarjopp]

After completing Braker+TSEBRA, I find that there are Busco genes called on the genome that are not present in the set of predicted amino acids. These Busco genes mostly are supported by my RNA-Seq reads. I would like to incorporate these genes in my final prediction set. What is the best way to do so? Do I re-run the pipeline and provide the Busco-generated AA sequences of the "missing" genes as protein evidence? If so, do I still provide the larger protein set I used initially? More fundamentally, how do I tell Braker to use the existing braker.gtf as a starting place, rather than just doing a de novo Braker run?

Also, if I wanted to increase the weight given to expression evidence, how could I do so? It is unclear to me why putative Busco genes that have mRNA support were not called as genes in the initial Braker run.

Many thanks, Sara

[tomasbruna]

Hi Sara,

for the first problem (using the existing BRAKER model and adding Busco-generated AA sequences of the "missing" genes as protein evidence) you can follow a procedure similar to the one in the https://github.com/Gaius-Augustus/BRAKER/issues/338 issue: Use the same species name,--skipAllTraining --hints TRAINING_WORKDIR/hintsfile.gff flags, and add the extra BUSCO proteins as --prot_seq busco.fa.

Also, if I wanted to increase the weight given to expression evidence, how could I do so? It is unclear to me why putative Busco genes that have mRNA support were not called as genes in the initial Braker run.

• You can increase the weight of particular hints by changing the src=E flag to src=M in the hintsfile (specified by --hints). All hints with the M source are enforced in the final prediction.
• If you want to increase the weight of all expression evidence, that can also be done, but the process is a bit more complicated. Let me know if you are interested in more details about this.

Sorry for the late response, hope you will still find the advice useful.

Best, Tomas