Comparative gene prediction in human and mouse

Abstract

The completion of the sequencing of the mouse genome promises to help predict human genes with greater accuracy. While current ab initio gene prediction programs are remarkably sensitive (i.e., they predict at least a fragment of most genes), their specificity is often low, predicting a large number of false-positive genes in the human genome. Sequence conservation at the protein level with the mouse genome can help eliminate some of those false positives. Here we describe SGP2, a gene prediction program that combines ab initio gene prediction with TBLASTX searches between two genome sequences to provide both sensitive and specific gene predictions. The accuracy of SGP2 when used to predict genes by comparing the human and mouse genomes is assessed on a number of data sets, including single-gene data sets, the highly curated human chromosome 22 predictions, and entire genome predictions from ENSEMBL. Results indicate that SGP2 outperforms purely ab initio gene prediction methods. Results also indicate that SGP2 works about as well with 3x shotgun data as it does with fully assembled genomes. SGP2 provides a high enough specificity that its predictions can be experimentally verified at a reasonable cost. SGP2 was used to generate a complete set of gene predictions on both the human and mouse by comparing the genomes of these two species. Our results suggest that another few thousand human and mouse genes currently not in ENSEMBL are worth verifying experimentally.

Figure 1.

Pairwise comparison using TBLASTX of the human and mouse genomic sequences coding for the HLA class II alpha chain. Black boxes indicate the coding exons, while black diagonals indicate the conserved alignments. The score of the conserved alignments (divided by 10) is given in the lower panels. Although conserved regions between the human and mouse genomic sequences coding for these genes fully include the coding exons, a substantial fraction of intronic regions is also conserved. The TBLASTX outptut was post-processed to show a continuous non-overlapping alignment.

Figure 2.

Rescoring of the exons predicted by GENEID according to the results of a TBLASTX search. See the “SGP2” section for a detailed explanation of the figure.

Figure 3.

Accuracy of the human and mouse SGP2 and GENSCAN predictions. The accuracy was measured in the entire chromosome sequences using the standard accuracy measures: SN, (sensitivity); SP, (specificity); CC, (correlation coefficient); SNe, (exon sensitivity); SPe, (exon specificity); and SNSP, (average of sensitivity and specificity at exon level). Predictions from both programs were compared against the human and mouse ENSEMBL annotations. Each dot corresponds to the accuracy measure of one chromosome. Chromosome labels are shown for outlier values. The boxplots (Tukey 1977) were obtained using the R-package (http://cran.r-project.org/).