Gene prediction programs

Bera TK, Lee S, Salvatore G, Lee B, Pastan I. MRP8, a new member of ABC transporter superfamily, identified by EST database mining and gene prediction program, is higly expressed in breast cancer. Mol Med 2001 7 509-516. 

Accuracy of gene prediction programs for single gene sequences from the [89] data set. Sn (sensitivity) is the number of exactly predicted exons/number of true exons (or nucleotides) Sp (specificity) is the number of exactly predicted exons/number of all predicted exons. Accuracy data for the programs developed before 1996 were estimated by Burset and Cuigo [89], The other data were received by the authors of the respective programs. 

Salamov A., Solovyev V. (1998) Fgenesh multiple gene prediction program httpd //genomic.sanger.ac.uk. 

Now we will generate an annotation of the D. ananassae genome using the SNAP (7) gene prediction program (wrapped by the runs nap script, Mercator distribution). The program zf f 2gtf (Mercator distribution) is used to convert from SNAP s ZFF format to GFF. 

The principal drawback of gene prediction programs is their inaccuracy. Not only do these programs overlook gene sequences whose structures do not conform to the rules applied in the training data, but they can also produce incorrect exon predictions. As a result, in our experience, intensive human quality control is required to review all predictions and eliminate these errors. 

---