GenoMiner: a tool for genome-wide search of coding and non-coding conserved sequence tags

Comparative evaluation of intron prediction methods and detection of plant genome annotation using intron length distributions

Intron prediction is an important problem of the constantly updated genome annotation. Using two model plant (rice and Arabidopsis) genomes, we compared two well-known intron prediction tools: the Blast-Like Alignment Tool (BLAT) and Sim4cc. The results showed that each of the tools had its own advantages and disadvantages. BLAT predicted more than 99% introns of whole genomic introns with a small number of false-positive introns. Sim4cc was successful at finding the correct introns with a false-negative rate of 1.02% to 4.85%, and it needed a longer run time than BLAT. Further, we evaluated the intron information of 10 complete plant genomes. As non-coding sequences, intron lengths are not limited by a triplet codon frame; so, intron lengths have three phases: a multiple of three bases (3n), a multiple of three bases plus one (3n + 1), and a multiple of three bases plus two (3n + 2). It was widely accepted that the percentages of the 3n, 3n + 1, and 3n + 2 introns were quite similar in genomes. Our studies showed that 80% (8/10) of species were similar in terms of the number of three phases. The percentages of 3n introns in Ostreococcus lucimarinus was excessive (47.7%), while in Ostreococcus tauri, it was deficient (29.1%). This discrepancy could have been the result of errors in intron prediction. It is suggested that a three-phase evaluation is a fast and effective method of detecting intron annotation problems.

Computational methods for ab initio and comparative gene finding

High-throughput DNA sequencing is increasing the amount of public complete genomes even though a precise gene catalogue for each organism is not yet available. In this context, computational gene finders play a key role in producing a first and cost-effective annotation. Nowadays a compilation of gene prediction tools has been made available to the scientific community and, despite the high number, they can be divided into two main categories: (1) ab initio and (2) evidence based. In the following, we will provide an overview of main methodologies to predict correct exon-intron structures of eukaryotic genes falling in such categories. We will take into account also new strategies that commonly refine ab initio predictions employing comparative genomics or other evidence such as expression data. Finally, we will briefly introduce metrics to in house evaluation of gene predictions in terms of sensitivity and specificity at nucleotide, exon, and gene levels as well.

Accurate discrimination of conserved coding and non-coding regions through multiple indicators of evolutionary dynamics

BACKGROUND

The conservation of sequences between related genomes has long been recognised as an indication of functional significance and recognition of sequence homology is one of the principal approaches used in the annotation of newly sequenced genomes. In the context of recent findings that the number non-coding transcripts in higher organisms is likely to be much higher than previously imagined, discrimination between conserved coding and non-coding sequences is a topic of considerable interest. Additionally, it should be considered desirable to discriminate between coding and non-coding conserved sequences without recourse to the use of sequence similarity searches of protein databases as such approaches exclude the identification of novel conserved proteins without characterized homologs and may be influenced by the presence in databases of sequences which are erroneously annotated as coding.

RESULTS

Here we present a machine learning-based approach for the discrimination of conserved coding sequences. Our method calculates various statistics related to the evolutionary dynamics of two aligned sequences. These features are considered by a Support Vector Machine which designates the alignment coding or non-coding with an associated probability score.

CONCLUSION

We show that our approach is both sensitive and accurate with respect to comparable methods and illustrate several situations in which it may be applied, including the identification of conserved coding regions in genome sequences and the discrimination of coding from non-coding cDNA sequences.

Statistical assessment of discriminative features for protein-coding and non coding cross-species conserved sequence elements

Background

The identification of protein coding elements in sets of mammalian conserved elements is one of the major challenges in the current molecular biology research. Many features have been proposed for automatically distinguishing coding and non coding conserved sequences, making so necessary a systematic statistical assessment of their differences. A comprehensive study should be composed of an association study, i.e. a comparison of the distributions of the features in the two classes, and a prediction study in which the prediction accuracies of classifiers trained on single and groups of features are analyzed, conditionally to the compared species and to the sequence lengths.

Results

In this paper we compared distributions of a set of comparative and non comparative features and evaluated the prediction accuracy of classifiers trained for discriminating sequence elements conserved among human, mouse and rat species. The association study showed that the analyzed features are statistically different in the two classes. In order to study the influence of the sequence lengths on the feature performances, a predictive study was performed on different data sets composed of coding and non coding alignments in equal number and equally long with an ascending average length. We found that the most discriminant feature was a comparative measure indicating the proportion of synonymous nucleotide substitutions per synonymous sites. Moreover, linear discriminant classifiers trained by using comparative features in general outperformed classifiers based on intrinsic ones. Finally, the prediction accuracy of classifiers trained on comparative features increased significantly by adding intrinsic features to the set of input variables, independently on sequence length (Kolmogorov-Smirnov P-value ≤ 0.05).

Conclusion

We observed distinct and consistent patterns for individual and combined use of comparative and intrinsic classifiers, both with respect to different lengths of sequences/alignments and with respect to error rates in the classification of coding and non-coding elements. In particular, we noted that comparative features tend to be more accurate in the classification of coding sequences – this is likely related to the fact that such features capture deviations from strictly neutral evolution expected as a consequence of the characteristics of the genetic code.

Genome-wide identification of coding and non-coding conserved sequence tags in human and mouse genomes

Background

The accurate detection of genes and the identification of functional regions is still an open issue in the annotation of genomic sequences. This problem affects new genomes but also those of very well studied organisms such as human and mouse where, despite the great efforts, the inventory of genes and regulatory regions is far from complete. Comparative genomics is an effective approach to address this problem. Unfortunately it is limited by the computational requirements needed to perform genome-wide comparisons and by the problem of discriminating between conserved coding and non-coding sequences. This discrimination is often based (thus dependent) on the availability of annotated proteins.

Results

In this paper we present the results of a comprehensive comparison of human and mouse genomes performed with a new high throughput grid-based system which allows the rapid detection of conserved sequences and accurate assessment of their coding potential. By detecting clusters of coding conserved sequences the system is also suitable to accurately identify potential gene loci.

Following this analysis we created a collection of human-mouse conserved sequence tags and carefully compared our results to reliable annotations in order to benchmark the reliability of our classifications. Strikingly we were able to detect several potential gene loci supported by EST sequences but not corresponding to as yet annotated genes.

Conclusion

Here we present a new system which allows comprehensive comparison of genomes to detect conserved coding and non-coding sequences and the identification of potential gene loci. Our system does not require the availability of any annotated sequence thus is suitable for the analysis of new or poorly annotated genomes.