Generalizations of Markov model to characterize biological sequences

iTriplet, a rule-based nucleic acid sequence motif finder

Background

With the advent of high throughput sequencing techniques, large amounts of sequencing data are readily available for analysis. Natural biological signals are intrinsically highly variable making their complete identification a computationally challenging problem. Many attempts in using statistical or combinatorial approaches have been made with great success in the past. However, identifying highly degenerate and long (>20 nucleotides) motifs still remains an unmet challenge as high degeneracy will diminish statistical significance of biological signals and increasing motif size will cause combinatorial explosion. In this report, we present a novel rule-based method that is focused on finding degenerate and long motifs. Our proposed method, named iTriplet, avoids costly enumeration present in existing combinatorial methods and is amenable to parallel processing.

Results

We have conducted a comprehensive assessment on the performance and sensitivity-specificity of iTriplet in analyzing artificial and real biological sequences in various genomic regions. The results show that iTriplet is able to solve challenging cases. Furthermore we have confirmed the utility of iTriplet by showing it accurately predicts polyA-site-related motifs using a dual Luciferase reporter assay.

Conclusion

iTriplet is a novel rule-based combinatorial or enumerative motif finding method that is able to process highly degenerate and long motifs that have resisted analysis by other methods. In addition, iTriplet is distinguished from other methods of the same family by its parallelizability, which allows it to leverage the power of today's readily available high-performance computing systems.

SpliceIT: a hybrid method for splice signal identification based on probabilistic and biological inference

Splice sites define the boundaries of exonic regions and dictate protein synthesis and function. The splicing mechanism involves complex interactions among positional and compositional features of different lengths. Computational modeling of the underlying constructive information is especially challenging, in order to decipher splicing-inducing elements and alternative splicing factors. SpliceIT (Splice Identification Technique) introduces a hybrid method for splice site prediction that couples probabilistic modeling with discriminative computational or experimental features inferred from published studies in two subsequent classification steps. The first step is undertaken by a Gaussian support vector machine (SVM) trained on the probabilistic profile that is extracted using two alternative position-dependent feature selection methods. In the second step, the extracted predictions are combined with known species-specific regulatory elements, in order to induce a tree-based modeling. The performance evaluation on human and Arabidopsis thaliana splice site datasets shows that SpliceIT is highly accurate compared to current state-of-the-art predictors in terms of the maximum sensitivity, specificity tradeoff without compromising space complexity and in a time-effective way. The source code and supplementary material are available at: http://www.med.auth.gr/research/spliceit/.

Variable-length positional modeling for biological sequence classification

Selecting the most informative features in supervised biological classification problems is a decisive pre-processing step for two main reasons: (1) to deal with the dimensionality reduction problem, and (2) to ascribe biological meaning to the underlying feature interactions. This paper presents a filter-based feature selection method that is suitable for positional modeling of biological sequences. The basic motivation is the problem of using a positional model of fixed length that sub-optimally describes biological sequences in a specific classification problem. The core filtering criterion is the F-score and the source features are the positional probabilities describing variable-length interactions among residues. The proposed method was evaluated on human splice sites classification using a linear SVM classifier. The method yields to superior classification accuracy compared to the individual positional models, while it maintains the space complexity of the individual models, in a time-efficient way and independently of the classifier.

MetaProm: a neural network based meta-predictor for alternative human promoter prediction

BACKGROUND

De novo eukaryotic promoter prediction is important for discovering novel genes and understanding gene regulation. In spite of the great advances made in the past decade, recent studies revealed that the overall performances of the current promoter prediction programs (PPPs) are still poor, and predictions made by individual PPPs do not overlap each other. Furthermore, most PPPs are trained and tested on the most-upstream promoters; their performances on alternative promoters have not been assessed.

RESULTS

In this paper, we evaluate the performances of current major promoter prediction programs (i.e., PSPA, FirstEF, McPromoter, DragonGSF, DragonPF, and FProm) using 42,536 distinct human gene promoters on a genome-wide scale, and with emphasis on alternative promoters. We describe an artificial neural network (ANN) based meta-predictor program that integrates predictions from the current PPPs and the predicted promoters' relation to CpG islands. Our specific analysis of recently discovered alternative promoters reveals that although only 41% of the 3' most promoters overlap a CpG island, 74% of 5' most promoters overlap a CpG island.

CONCLUSION

Our assessment of six PPPs on 1.06 x 109 bps of human genome sequence reveals the specific strengths and weaknesses of individual PPPs. Our meta-predictor outperforms any individual PPP in sensitivity and specificity. Furthermore, we discovered that the 5' alternative promoters are more likely to be associated with a CpG island.

Search for basonuclin target genes

Basonuclin (Bnc 1) is a transcription factor that has an unusual ability to interact with promoters of both RNA polymerases I and II. The action of basonuclin is mediated through three pairs of evolutionarily conserved zinc fingers, which produce three DNase I footprints on the promoters of rDNA and the basonuclin gene. Using these DNase footprints, we built a computational model for the basonuclin DNA-binding module, which was used to identify in silico potential RNA polymerase II target genes in the human and mouse promoter databases. The target genes of basonuclin show that it regulates the expression of proteins involved in chromatin structure, transcription/DNA-binding, ion-channels, adhesion/cell-cell junction, signal transduction, and intracellular transport. Our results suggest that basonuclin, like MYC, may coordinate transcriptional activities among the three RNA polymerases. But basonuclin regulates a distinctive set of pathways, which differ from that regulated by MYC.

A mammalian promoter model links cis elements to genetic networks

An accurate identification of gene promoters remains an important challenge. Computational approaches for this problem rely on promoter sequence attributes that are believed to be critical for transcription initiation. Here we report a probabilistic model that captures two important properties of promoters, not used by previous methods, viz., the location preference and co-occurrence of promoter elements. Additionally, we found that many of the position-specific DNA elements are strongly linked with the function of the gene product. For instance, a highly conserved motif CCTTT at -1 position is strongly associated with protein synthesis, cellular and tissue development. Our comparative analysis of promoter classes reveals that the promoters devoid of CpG islands are more conserved and have fewer alternative transcription start sites. The discovered links between promoter elements and gene function allows us to infer genetic networks from promoter elements. The web server for the PSPA promoter predictor is available at /PSPA.