funannotate predict

[mictadlo]

Hi, I do not plan to use GeneMark-ET but I want to use Augustus, GlimmerHMM and SNAP.

1. If I ran funannotate train do I still have to use funannotate predict --rna_bam? If yes then what is the difference between train and predict?
2. Is Is possible to use the output from Mikado as the --stringtie parameter? This tool apparently shows

   ..that the accuracy of transcript reconstruction can be boosted by combining multiple methods, and we present a novel algorithm to integrate multiple RNA-seq assemblies into a coherent transcript annotation. Our algorithm can remove redundancies and select the best transcript models according to user-specified metrics while solving common artifacts such as erroneous transcript chimerisms.

Thank you in advance,

Michal

[hyphaltip]

For 1. No - predict will find the RNA BAM files automatically if you ran train already. note that train will also run full PASA to get polished spliced sites and models for SNAP and augustus training which require a set of gene models in gff/gbk.

[mictadlo]

Thank you, is Is possible to use the output from Mikado as the --stringtie parameter? This tool apparently shows

..that the accuracy of transcript reconstruction can be boosted by combining multiple methods, and we present a novel algorithm to integrate multiple RNA-seq assemblies into a coherent transcript annotation. Our algorithm can remove redundancies and select the best transcript models according to user-specified metrics while solving common artifacts such as erroneous transcript chimerisms.

[hyphaltip]

any assembled transcriptome is an option but I think you still need to give left and right reads for it to select genes with minimal expression level for training step.