AUGUSTUS is a gene prediction program for eukaryotes genomic sequences. It can be used as an ab initio program, which means it bases its prediction purely on the sequence. AUGUSTUS may also incorporate hints on the gene structure coming from extrinsic sources such as EST, MS or MS, protein alignments and synthenic genomic alignments. AUGUSTUS usually belongs to the most accurate programs for the species it is trained for. Often it is the most accurate ab initio program. For example, at the independent gene finder assessment (EGASP) on the human ENCODE regions AUGUSTUS was the most accurate gene finder among the tested ab initio programs. At the more recent nGASP (worm), it was among the best in the ab initio and transcript-based categories. Gene prediction accuracy of AUGUSTUS always depends on the quality of the training gene set that was used for training species specific parameters. You should not expect the greatest accuracy from fully automated training gene generation as provided by this web server application. Instead, you should manually inspect (and maybe interatively improve) the training gene set.
Parent program: augustus
AUGUSTUS is a software to predict genes in eukaryotic genomic sequences.The approach is based on a generalized Hidden Markov Model. AUGUSTUS parses a genomic sequence to segments that are most likely according to the underlying statistical models. The program is often the most accurate for ab initio gene prediction for the species it is trained for. AUGUSTUS is able to incorporate EST, mRNA and protein alignments. It also can predict alternative transcripts, splicing and 5'UTR and 3'UTR including introns with probabilities for the transcripts and each exon and intron.