Abstract
A common practice in computational genomic analysis is to use a set of ‘background’ sequences as negative controls for evaluating the false-positive rates of prediction tools, such as gene identification programs and algorithms for detection of cis-regulatory elements. Such ‘background’ sequences are generally taken from regions of the genome presumed to be intergenic, or generated synthetically by ‘shuffling’ real sequences. This last method can lead to underestimation of false-positive rates. We developed a new method for generating artificial sequences that are modeled after real intergenic sequences in terms of composition, complexity and interspersed repeat content. These artificial sequences can serve as an inexhaustible source of high-quality negative controls. We used artificial sequences to evaluate the false-positive rates of a set of programs for detecting interspersed repeats, ab initio prediction of coding genes, transcribed regions and non-coding genes. We found that RepeatMasker is more accurate than PClouds, Augustus has the lowest false-positive rate of the coding gene prediction programs tested, and Infernal has a low false-positive rate for non-coding gene detection. A web service, source code and the models for human and many other species are freely available at http://repeatmasker.org/garlic/.
Highlights
Genomes evolve by random accumulation of mutations and by selection for a variety of functional requirements
We model the genome as being composed of three classes of sequence: (i) sequences under functional or mutational constraints, (ii) sequences that arose by duplication but are largely unconstrained, and (iii) a background or ‘base’ sequence
Based on the available annotation of the human genome, we identified 574 Mb of ‘base’ sequence (17% of the genome) left after removing all fragments annotated as genes, pseudogenes, CpG islands, ultraconserved sequences and repetitive sequences
Summary
Genomes evolve by random accumulation of mutations and by selection for a variety of functional requirements. For species with short generation time and large population sizes (e.g. bacteria), the strong selective forces lead to highly optimized genomes, dense in genes and with negligible overhead of non-functional sequences: this makes prokaryotic gene prediction relatively straightforward [1]. The genomes of species with much longer generation times and much reduced population sizes (e.g. vertebrates) accumulate vast amounts of genetic material that largely appears not to be under selective constraints [2]. Functional sequences and regulatory elements are a small fraction of the vertebrate genome, making their identification difficult. Recognizing alternative splicing demands further algorithmic complexity, as does modeling of non-coding transcripts. For all these reasons and more, ab initio vertebrate gene prediction poses a significant challenge for computational biology
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
