v2.1.5

Novel features of BRAKER in v2.1.5 (compared to 2.1.4):

• integration of protein evidence from ProtHint with GeneMark-EP/EP+; this should now be the method of choice for running BRAKER with protein evidence since it does not require prior knowledge about how closely or distantly the proteins are to the target species

• integration of protein evidence from ProtHint and RNA-Seq evidence from short read to genome alignment with GeneMark-ETP (development version); this feature will in the future continue to undergo improvement

• novel output file braker.gtf that contains all transcript variants predicted by AUGUSTUS and GeneMark-EX

• output of what-to-cite.txt that lists which publications should be cited depending on an individual BRAKER run

v2.1.4

Novel features of BRAKER in v2.1.4 (compared to 2.1.2):

• fully automated generation of track data hubs for display of BRAKER results with the UCSC Genome Browser using MakeHub (https://github.com/Gaius-Augustus/MakeHub)

• random reduction of training data for AUGUSTUS if training set is bigger than 8000 genes

• deletion of files that are typically not needed after a BRAKER run has finished successfully

• improved integration of GeneMark-EX, e.g. new expert options --transmasked_fasta and --min_contig

• improved integration of GeneMark-EX and AUGUSTUS in ETP mode (RNA-Seq & proteins as evidence)

• option to replace NCBI BLAST by DIAMOND for speeding up a BRAKER run

• migration from bam2wig binary (that used to cause compilation issues on many machines) to bamToWig.py

• fixing of AUGUSTUS genes with spliced (= in frame) stop codons

• --softwareCheck option if you don't want to run BRAKER but only check whether required software for your run is present

v2.1.2

This release of BRAKER includes the novel feature of automatic generation of codingseq and protein FASTA-files from AUGUSTUS GTF-predictions in all running modes.