Improving the Caenorhabditis elegans genome annotation using machine learning

For modern biology, precise genome annotations are of prime importance, as they allow the accurate definition of genic regions. We employ state-of-the-art machine learning methods to assay and improve the accuracy of the genome annotation of the nematode Caenorhabditis elegans. The proposed machine learning system is trained to recognize exons and introns on the unspliced mRNA, utilizing recent advances in support vector machines and label sequence learning. In 87% (coding and untranslated regions) and 95% (coding regions only) of all genes tested in several out-of-sample evaluations, our method correctly identified all exons and introns. Notably, only 37% and 50%, respectively, of the presently unconfirmed genes in the C. elegans genome annotation agree with our predictions, thus we hypothesize that a sizable fraction of those genes are not correctly annotated. A retrospective evaluation of the Wormbase WS120 annotation [] of C. elegans reveals that splice form predictions on unconfirmed genes in WS120 are inaccurate in about 18% of the considered cases, while our predictions deviate from the truth only in 10%-13%. We experimentally analyzed 20 controversial genes on which our system and the annotation disagree, confirming the superiority of our predictions. While our method correctly predicted 75% of those cases, the standard annotation was never completely correct. The accuracy of our system is further corroborated by a comparison with two other recently proposed systems that can be used for splice form prediction: SNAP and ExonHunter. We conclude that the genome annotation of C. elegans and other organisms can be greatly enhanced using modern machine learning technology.     

利用不同G+C含量细菌基因组评估细菌ncRNA基因预测工具

【目的】为识别已完成全测序细菌基因组中的ncRNA基因,对3个常用ncRNA预测工具s RNAPredict、PORTRAIT和s RNAscanner进行评估。【方法】选择了细菌ncRNA数据库BSRD收录的含有已知ncRNA基因数目大于30的9个细菌基因组,并按基因组G+C含量进行分类,比较s RNAPredict和PORTRAIT工具的预测准确性。提取不同G+C含量基因组中ncRNA基因转录起始和终止区的序列特征,对s RNAscanner预测结果进行评估。【结果】s RNAPredict对细菌ncRNA基因的预测特异性和阳性检出率均高于PORTRAIT,而敏感性则较差;两种工具预测效果均随基因组G+C含量不同而产生明显变化。在不同G+C含量的细菌基因组中,ncRNA基因启动子和终止子区域的序列特征有明显差异。利用这些序列特征能提高s RNAscanner预测ncRNA基因的平均水平。【结论】3种ncRNA基因工具预测效果随基因组G+C含量变化而不同。不同G+C含量基因组中ncRNA基因的转录起始和终止区特征可作为ncRNA基因预测的重要参数之一。     

Ensemble Positive Unlabeled Learning for Disease Gene Identification

An increasing number of genes have been experimentally confirmed in recent years as causative genes to various human diseases. The newly available knowledge can be exploited by machine learning methods to discover additional unknown genes that are likely to be associated with diseases. In particular, positive unlabeled learning (PU learning) methods, which require only a positive training set P (confirmed disease genes) and an unlabeled set U (the unknown candidate genes) instead of a negative training set N, have been shown to be effective in uncovering new disease genes in the current scenario. Using only a single source of data for prediction can be susceptible to bias due to incompleteness and noise in the genomic data and a single machine learning predictor prone to bias caused by inherent limitations of individual methods. In this paper, we propose an effective PU learning framework that integrates multiple biological data sources and an ensemble of powerful machine learning classifiers for disease gene identification. Our proposed method integrates data from multiple biological sources for training PU learning classifiers. A novel ensemble-based PU learning method EPU is then used to integrate multiple PU learning classifiers to achieve accurate and robust disease gene predictions. Our evaluation experiments across six disease groups showed that EPU achieved significantly better results compared with various state-of-the-art prediction methods as well as ensemble learning classifiers. Through integrating multiple biological data sources for training and the outputs of an ensemble of PU learning classifiers for prediction, we are able to minimize the potential bias and errors in individual data sources and machine learning algorithms to achieve more accurate and robust disease gene predictions. In the future, our EPU method provides an effective framework to integrate the additional biological and computational resources for better disease gene predictions.     

Automatic prediction of non-coding RNA genes in prokaryotes based on compositional statistics

Although non-coding RNA (ncRNA) genes do not encode proteins, they play vital roles in cells by producing functionally important RNAs. In this paper, we present a novel method for predicting ncRNA genes based on compositional features extracted directly from gene sequences. Our method consists of two Support Vector Machine (SVM) models--Codon model which uses codon usage features derived from ncRNA genes and protein-coding genes and Kmer model which utilizes features of nucleotide and dinucleotide frequency extracted respectively from ncRNA genes and randomly chosen genome sequences. The 10-fold cross-validation accuracy for the two models is found to be 92% and 91%, respectively. Thus, we could make an automatic prediction of ncRNA genes in one genome without manual filtration of protein-coding genes. After applying our method in Sulfolobus solfataricus genome, 25 prediction results have been generated according to 25 cut-off pairs. We have also applied the approach in E. coli and found our results comparable to those of previous studies. In general, our method enables automatic identification of ncRNA genes in newly sequenced prokaryotic genomes.     

Evolutionary Modeling and Prediction of Non-Coding RNAs in Drosophila

We performed benchmarks of phylogenetic grammar-based ncRNA gene prediction, experimenting with eight different models of structural evolution and two different programs for genome alignment. We evaluated our models using alignments of twelve Drosophila genomes. We find that ncRNA prediction performance can vary greatly between different gene predictors and subfamilies of ncRNA gene. Our estimates for false positive rates are based on simulations which preserve local islands of conservation; using these simulations, we predict a higher rate of false positives than previous computational ncRNA screens have reported. Using one of the tested prediction grammars, we provide an updated set of ncRNA predictions for D. melanogaster and compare them to previously-published predictions and experimental data. Many of our predictions show correlations with protein-coding genes. We found significant depletion of intergenic predictions near the 3′ end of coding regions and furthermore depletion of predictions in the first intron of protein-coding genes. Some of our predictions are colocated with larger putative unannotated genes: for example, 17 of our predictions showing homology to the RFAM family snoR28 appear in a tandem array on the X chromosome; the 4.5 Kbp spanned by the predicted tandem array is contained within a FlyBase-annotated cDNA.     

A computational approach to identify genes for functional RNAs in genomic sequences

Currently there is no successful computational approach for identification of genes encoding novel functional RNAs (fRNAs) in genomic sequences. We have developed a machine learning approach using neural networks and support vector machines to extract common features among known RNAs for prediction of new RNA genes in the unannotated regions of prokaryotic and archaeal genomes. The Escherichia coli genome was used for development, but we have applied this method to several other bacterial and archaeal genomes. Networks based on nucleotide composition were 80–90% accurate in jackknife testing experiments for bacteria and 90–99% for hyperthermophilic archaea. We also achieved a significant improvement in accuracy by combining these predictions with those obtained using a second set of parameters consisting of known RNA sequence motifs and the calculated free energy of folding. Several known fRNAs not included in the training datasets were identified as well as several hundred predicted novel RNAs. These studies indicate that there are many unidentified RNAs in simple genomes that can be predicted computationally as a precursor to experimental study. Public access to our RNA gene predictions and an interface for user predictions is available via the web.     

SCGPred： A Score-based Method for Gene Structure Prediction by Combining Multiple Sources of Evidence

Predicting protein-coding genes still remains a significant challenge. Although a variety of computational programs that use commonly machine learning methods have emerged, the accuracy of predictions remains a low level when implementing in large genomic sequences. Moreover, computational gene finding in newly se- quenced genomes is especially a difficult task due to the absence of a training set of abundant validated genes. Here we present a new gene-finding program, SCGPred, to improve the accuracy of prediction by combining multiple sources of evidence. SCGPred can perform both supervised method in previously well-studied genomes and unsupervised one in novel genomes. By testing with datasets composed of large DNA sequences from human and a novel genome of Ustilago maydi, SCGPred gains a significant improvement in comparison to the popular ab initio gene predictors. We also demonstrate that SCGPred can significantly improve prediction in novel genomes by combining several foreign gene finders with similarity alignments, which is superior to other unsupervised methods. Therefore, SCGPred can serve as an alternative gene-finding tool for newly sequenced eukaryotic genomes. The program is freely available at http：//bio.scu.edu.cn/SCGPred/.     

PROSO II--a new method for protein solubility prediction

Many fields of science and industry depend on efficient production of active protein using heterologous expression in Escherichia coli. The solubility of proteins upon expression is dependent on their amino acid sequence. Prediction of solubility from sequence is therefore highly valuable. We present a novel machine-learning-based model called PROSO II which makes use of new classification methods and growth in experimental data to improve coverage and accuracy of solubility predictions. The classification algorithm is organized as a two-layered structure in which the output of a primary Parzen window model for sequence similarity and a logistic regression classifier of amino acid k-mer composition serve as input for a second-level logistic regression classifier. Compared with previously published research our model is trained on five times more data than used by any other method before (82 000 proteins). When tested on a separate holdout set not used at any point of method development our server attained the best results in comparison with other currently available methods: accuracy 75.4%, Matthew's correlation coefficient 0.39, sensitivity 0.731, specificity 0.759, gain (soluble) 2.263. In summary, due to utilization of cutting edge machine learning technologies combined with the largest currently available experimental data set the PROSO II server constitutes a substantial improvement in protein solubility predictions. PROSO II is available at http://mips.helmholtz-muenchen.de/prosoII.     

A novel method of protein secondary structure prediction with high segment overlap measure: support vector machine approach

We have introduced a new method of protein secondary structure prediction which is based on the theory of support vector machine (SVM). SVM represents a new approach to supervised pattern classification which has been successfully applied to a wide range of pattern recognition problems, including object recognition, speaker identification, gene function prediction with microarray expression profile, etc. In these cases, the performance of SVM either matches or is significantly better than that of traditional machine learning approaches, including neural networks.The first use of the SVM approach to predict protein secondary structure is described here. Unlike the previous studies, we first constructed several binary classifiers, then assembled a tertiary classifier for three secondary structure states (helix, sheet and coil) based on these binary classifiers. The SVM method achieved a good performance of segment overlap accuracy SOV=76.2 % through sevenfold cross validation on a database of 513 non-homologous protein chains with multiple sequence alignments, which out-performs existing methods. Meanwhile three-state overall per-residue accuracy Q(3) achieved 73.5 %, which is at least comparable to existing single prediction methods. Furthermore a useful "reliability index" for the predictions was developed. In addition, SVM has many attractive features, including effective avoidance of overfitting, the ability to handle large feature spaces, information condensing of the given data set, etc. The SVM method is conveniently applied to many other pattern classification tasks in biology.     

LocARNA-P: accurate boundary prediction and improved detection of structural RNAs

Current genomic screens for noncoding RNAs (ncRNAs) predict a large number of genomic regions containing potential structural ncRNAs. The analysis of these data requires highly accurate prediction of ncRNA boundaries and discrimination of promising candidate ncRNAs from weak predictions. Existing methods struggle with these goals because they rely on sequence-based multiple sequence alignments, which regularly misalign RNA structure and therefore do not support identification of structural similarities. To overcome this limitation, we compute columnwise and global reliabilities of alignments based on sequence and structure similarity; we refer to these structure-based alignment reliabilities as STARs. The columnwise STARs of alignments, or STAR profiles, provide a versatile tool for the manual and automatic analysis of ncRNAs. In particular, we improve the boundary prediction of the widely used ncRNA gene finder RNAz by a factor of 3 from a median deviation of 47 to 13 nt. Post-processing RNAz predictions, LocARNA-P's STAR score allows much stronger discrimination between true- and false-positive predictions than RNAz's own evaluation. The improved accuracy, in this scenario increased from AUC 0.71 to AUC 0.87, significantly reduces the cost of successive analysis steps. The ready-to-use software tool LocARNA-P produces structure-based multiple RNA alignments with associated columnwise STARs and predicts ncRNA boundaries. We provide additional results, a web server for LocARNA/LocARNA-P, and the software package, including documentation and a pipeline for refining screens for structural ncRNA, at http://www.bioinf.uni-freiburg.de/Supplements/LocARNA-P/.     

RNAstrand: reading direction of structured RNAs in multiple sequence alignments

Motivation

Genome-wide screens for structured ncRNA genes in mammals, urochordates, and nematodes have predicted thousands of putative ncRNA genes and other structured RNA motifs. A prerequisite for their functional annotation is to determine the reading direction with high precision.

Results

While folding energies of an RNA and its reverse complement are similar, the differences are sufficient at least in conjunction with substitution patterns to discriminate between structured RNAs and their complements. We present here a support vector machine that reliably classifies the reading direction of a structured RNA from a multiple sequence alignment and provides a considerable improvement in classification accuracy over previous approaches.

Software

RNAstrand is freely available as a stand-alone tool from http://www.bioinf.uni-leipzig.de/Software/RNAstrand and is also included in the latest release of RNAz, a part of the Vienna RNA Package.     

Inferring adaptive regulation thresholds and association rules from gene expression data through combinatorial optimization learning

There is a need to design computational methods to support the prediction of gene regulatory networks. Such models should offer both biologically-meaningful and computationally-accurate predictions, which in combination with other techniques may improve large-scale, integrative studies. This paper presents a new machine learning method for the prediction of putative regulatory associations from expression data, which exhibit properties never or only partially addressed by other techniques recently published. The method was tested on a Saccharomyces cerevisiae gene expression dataset. The results were statistically validated and compared with the relationships inferred by two machine learning approaches to gene regulatory network prediction. Furthermore, the resulting predictions were assessed using domain knowledge. The proposed algorithm may be able to accurately predict relevant biological associations between genes. One of the most relevant features of this new method is the prediction of adaptive regulation thresholds for the discretization of gene expression values, which is required prior to the rule association learning process. Moreover, an important advantage consists of its low computational cost to infer association rules. The proposed system may significantly support exploratory, large-scale studies of automated identification of potentially-relevant gene expression associations.