MAKER - How to set weight for merge legacy annotations

Question

[1300189805]

Good day. I am new to genome annotation.I am running maker to merge evidence from est, homolog, augustus and braker.
The following is maker_opts.ctl:

genome=genome.fasta
est_gff=transcript.gff3
protein=homolog.gffs
pred_gff=Augustus.gff3, Braker.gff3

I have executed the "mpiexec -n 30 maker > maker.out " command with the default configuration file and got 'genome.all.gff file', but the busco evaluation completeness is very low. My question is next how should I adjust the weight value of each evidence to get better annotation results. I want to know which parameters or configuration files should be adjusted? Thank you. Will appreciate any help

[Juke34]

What is the result of your BUSCO annotation completeness? And what as the assembly one? Have a look at these gists:

• first one (and the associated course)
• second one

It might help.

If your annotation is fragmented you will have to play on max_dna_length and split_hitparameters. If there is lot of missing gene, you can verify that you keep pure abinitio prediction and not only prediction supported by evidences (see parameter keep_preds=0 #Add unsupported gene prediction to final annotation set, 1 = yes, 0 = no)

Check yo have activated pure evidence base prediction too :

est2genome=0 #infer gene predictions directly from ESTs, 1 = yes, 0 = no
protein2genome=0 #gene prediction from protein homology, 1 = yes, 0 = no

The best is also to visualize. your results in a genome browser, that help to determine what could be the problem. If the abinitio prediction are not good, you will have to train properly your abinitio before to re-run a MAKER with the new abinitio profile.