The programs usage in Scientific publications
FGENESH+ - HMM plus similar protein-based gene prediction [Help] [Example]
Speed and Accuracy of Fgenesh+
PROT_MAP - mapping of a set of proteins on genome [Help] [Example]
FGENESH_C - HMM plus similar cDNA-based gene structure prediction (multiple genes, both chains) [Help] [Example]
FGENESH-2 - HMM gene prediction using two genomic sequences of close organisms (as Human and Mouse) [Help] [Example]
FGENESH+ and FGENESH_C programs can be used if there is a protein or cDNA/EST sequence similar to that of predicted gene. For example, you can run ab initio gene finding programs as FGENES or FGENESH and run BLASTP DB search with the predicted exons. Any true predicted exon can provide you with known similar protein, if such protein exists in DB. Take sequence of homologous protein and run FGENESH+. The accuracy of gene prediction can be up to 100%, depending on how similar the predicted and DB proteins are.
FGENESH-2 program can be used if there are sequences from two related organisms available, such as human and mouse. The program gives higher score to exons that have predicted amino acid sequences homologous to that of related organism's exons, which allows to substantially more accurate exon prediction and gene assembly.
FGENESH++ is our newest gene prediction program that works in following steps: (1) performs ab initio gene prediction using FGENESH algorithm; (2) runs predicted amino acid sequences of all potential exons through NR protein sequence database using DBSCAN-P engine; and (3) runs second round of gene prediction with higher scores assigned to exons homologous to known proteins.
The result is fully automated genome annotation of quality similar to manual annotation. The program is extremely fast - whole human genome is annotated in 20-30 hours on a machine like 500 MHz DEC Alpha. Due to computer resources constraints, we cannot make FGENESH++ available through the web site - you have to use FGENESH+ instead.