Softberry completed fully automatic annotation of Human genome by FGENESH++ program. Of 38,570 predicted genes, 27,851 (72%) have homologs in NR protein database. Results are presented for public in the Genome Browser of UCSC Human Genome Project Team site
Fgenesh++ predictions are based on Softberry's gene finding software. Fgenesh++ uses both HMMs and protein similarity to find genes in a completely automated manner.
Fgenesh++ uses FGENESH and FGENESH+ programs. Fgenesh is ab initio gene prediction works very fast processing Human genome sequences for about 4 hours.
The programs also have parameters for many model organisms such as Drosophila,
Yeast, Dicot and Monocot plants.
Later we plan to annotate also CELL LOCATION of predicted proteins
(by our ProtComp program available at:http://www.softberry.com/protein.html)