PSoL: a positive sample only learning algorithm for finding non-coding RNA genes

MOTIVATION: Small non-coding RNA (ncRNA) genes play important regulatory roles in a variety of cellular processes. However, detection of ncRNA genes is a great challenge to both experimental and computational approaches. In this study, we describe a new approach called positive sample only learning (PSoL) to predict ncRNA genes in the Escherichia coli genome. Although PSoL is a machine learning method for classification, it requires no negative training data, which, in general, is hard to define properly and affects the performance of machine learning dramatically. In addition, using the support vector machine (SVM) as the core learning algorithm, PSoL can integrate many different kinds of information to improve the accuracy of prediction. Besides the application of PSoL for predicting ncRNAs, PSoL is applicable to many other bioinformatics problems as well. RESULTS: The PSoL method is assessed by 5-fold cross-validation experiments which show that PSoL can achieve about 80% accuracy in recovery of known ncRNAs. We compared PSoL predictions with five previously published results. The PSoL method has the highest percentage of predictions overlapping with those from other methods.     

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.     

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.     

Systematic identification and characterization of chicken (Gallus gallus) ncRNAs

Recent studies have demonstrated that non-coding RNAs (ncRNAs) play important roles during development and evolution. Chicken, the first genome-sequenced non-mammalian amniote, possesses unique features for developmental and evolutionary studies. However, apart from microRNAs, information on chicken ncRNAs has mainly been obtained from computational predictions without experimental validation. In the present study, we performed a systematic identification of intermediate size ncRNAs (50–500 nt) by ncRNA library construction and identified 125 chicken ncRNAs. Importantly, through the bioinformatics and expression analysis, we found the chicken ncRNAs has several novel features: (i) comparative genomic analysis against 18 sequenced vertebrate genomes revealed that the majority of the newly identified ncRNA candidates is not conserved and most are potentially bird/chicken specific, suggesting that ncRNAs play roles in lineage/species specification during evolution. (ii) The expression pattern analysis of intronic snoRNAs and their host genes suggested the coordinated expression between snoRNAs and their host genes. (iii) Several spatio-temporal specific expression patterns suggest involvement of ncRNAs in tissue development. Together, these findings provide new clues for future functional study of ncRNAs during development and evolution.     

Generation of a neuro-specific microarray reveals novel differentially expressed noncoding RNAs in mouse models for neurodegenerative diseases

We have generated a novel, neuro-specific ncRNA microarray, covering 1472 ncRNA species, to investigate their expression in different mouse models for central nervous system diseases. Thereby, we analyzed ncRNA expression in two mouse models with impaired calcium channel activity, implicated in Epilepsy or Parkinson''s disease, respectively, as well as in a mouse model mimicking pathophysiological aspects of Alzheimer''s disease. We identified well over a hundred differentially expressed ncRNAs, either from known classes of ncRNAs, such as miRNAs or snoRNAs or which represented entirely novel ncRNA species. Several differentially expressed ncRNAs in the calcium channel mouse models were assigned as miRNAs and target genes involved in calcium signaling, thus suggesting feedback regulation of miRNAs by calcium signaling. In the Alzheimer mouse model, we identified two snoRNAs, whose expression was deregulated prior to amyloid plaque formation. Interestingly, the presence of snoRNAs could be detected in cerebral spine fluid samples in humans, thus potentially serving as early diagnostic markers for Alzheimer''s disease. In addition to known ncRNAs species, we also identified 63 differentially expressed, entirely novel ncRNA candidates, located in intronic or intergenic regions of the mouse genome, genomic locations, which previously have been shown to harbor the majority of functional ncRNAs.     

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/.     

Plant noncoding RNA gene discovery by "single-genome comparative genomics"

Plant genomes have undergone multiple rounds of duplications that contributed massively to the growth of gene families. The structure of resulting families has been studied in depth for protein-coding genes. However, little is known about the impact of duplications on noncoding RNA (ncRNA) genes. Here we perform a systematic analysis of duplicated regions in the rice genome in search of such ncRNA repeats. We observe that, just like their protein counterparts, most ncRNA genes have undergone multiple duplications that left visible sequence conservation footprints. The extent of ncRNA gene duplication in plants is such that these sequence footprints can be exploited for the discovery of novel ncRNA gene families on a large scale. We developed an SVM model that is able to retrieve likely ncRNA candidates among the 100,000+ repeat families in the rice genome, with a reasonably low false-positive discovery rate. Among the nearly 4000 ncRNA families predicted by this means, only 90 correspond to putative snoRNA or miRNA families. About half of the remaining families are classified as structured RNAs. New candidate ncRNAs are particularly enriched in UTR and intronic regions. Interestingly, 89% of the putative ncRNA families do not produce a detectable signal when their sequences are compared to another grass genome such as maize. Our results show that a large fraction of rice ncRNA genes are present in multiple copies and are species-specific or of recent origin. Intragenome comparison is a unique and potent source for the computational annotation of this major class of ncRNA.     

Identification of small non-coding RNAs from mitochondria and chloroplasts

Small non-protein-coding RNAs (ncRNAs) have been identified in a wide spectrum of organisms ranging from bacteria to humans. In eukarya, systematic searches for ncRNAs have so far been restricted to the nuclear or cytosolic compartments of cells. Whether or not small stable non-coding RNA species also exist in cell organelles, in addition to tRNAs or ribosomal RNAs, is unknown. We have thus generated cDNA libraries from size-selected mammalian mitochondrial RNA and plant chloroplast RNA and searched for small ncRNA species in these two types of DNA-containing cell organelles. In total, we have identified 18 novel candidates for organellar ncRNAs in these two cellular compartments and confirmed expression of six of them by northern blot analysis or RNase A protection assays. Most candidate ncRNA genes map to intergenic regions of the organellar genomes. As found previously in bacteria, the presumptive ancestors of present-day chloroplasts and mitochondria, we also observed examples of antisense ncRNAs that potentially could target organelle-encoded mRNAs. The structural features of the identified ncRNAs as well as their possible cellular functions are discussed. The absence from our libraries of abundant small RNA species that are not encoded by the organellar genomes suggests that the import of RNAs into cell organelles is of very limited significance or does not occur at all.     

Codon bias and base composition are poor indicators of horizontally transferred genes

Horizontal gene transfer is now recognized as an important mechanism of evolution. Several methods to detect horizontally transferred genes have been suggested. These methods are based on either nucleotide composition or the failure to find a similar gene in closely related species. Genes that evolve vertically between closely related species can be divided into those that retain homologous chromosomal positions (positional orthologs) and those that do not. By comparing open reading frames in the Escherichia coli and Salmonella typhi genomes, we identified 2,728 positional orthologs since these species split 100 MYA. A group of 1,144 novel E. coli genes were unusually diverged from their S. typhi counterparts. These novel genes included those that had been horizontally transferred into E. coli, as well as members of gene pairs that had been rearranged or deleted. Positional orthologs were used to investigate compositional methods of identifying horizontally transferred genes. A large number of E. coli genes with normal nucleotide composition have no apparent ortholog in S. typhi, and many genes of atypical composition do, in fact, have positional orthologs. A phylogenetic approach was employed to confirm selected examples of horizontal transmission among the novel groups of genes. Our analysis of 80 E. coli genes determined that a number of genes previously classified as horizontally transferred based on base composition and codon bias were native, and genes previously classified as native appeared to be horizontally transferred. Hence, atypical nucleotide composition alone is not a reliable indicator of horizontal transmission.     

Rare codon clusters at 5'-end influence heterologous expression of archaeal gene in Escherichia coli

Proteins from hyperthermophilic microorganisms are attractive candidates for novel biocatalysts because of their high resistance to temperature extremes. However, archaeal genes are usually poorly expressed in Escherichia coli because of differences in codon usage. Genes from the thermoacidophilic archaea Sulfolobus solfataricus and Thermoplasma acidophilum contain high proportions of rare codons for arginine, isoleucine, and leucine, which are recognized by the tRNAs encoded by the argU, ileY, and leuW genes, respectively, and which are rarely used in E. coli. To examine the effects of these rare codons on heterologous expression, we expressed the Sso_gnaD and Tac_gnaD genes from S. solfataricus and T. acidophilum, respectively, in E. coli. The Sso_gnaD product was expressed at very low levels when the open reading frame (ORF) was cloned in pRSET and expressed in E. coli BL21(DE3), and was expressed at much higher levels in the E. coli BL21(DE3)-CodonPlus RIL strain, which contains extra copies of the argU, ileY, and leuW tRNA genes. In contrast, Tac_gnaD was expressed at similar levels in both E. coli strains. Comparison of the Sso_gnaD and Tac_gnaD gene sequences revealed that the 5'-end of the Sso_gnaD sequence was rich in AGA(arg) and ATA(Ile) codons. These codons were replaced with the codons commonly used in E. coli by polymerase chain reaction-mediated site-directed mutagenesis. The results of expression studies showed that a non-tandem repeat of rare codons is critical in the observed interference in heterologous expression of this gene. We concluded that the level of heterologous expression of Sso_gnaD in E. coli was limited by the clustering of the rare codons in the ORF, rather than on the rare codon frequency.     

Bridging consent: from toll bridges to lift bridges?

Background

One of the best and most accurate methods for identifying disease-causing genes is monitoring gene expression values in different samples using microarray technology. One of the shortcomings of microarray data is that they provide a small quantity of samples with respect to the number of genes. This problem reduces the classification accuracy of the methods, so gene selection is essential to improve the predictive accuracy and to identify potential marker genes for a disease. Among numerous existing methods for gene selection, support vector machine-based recursive feature elimination (SVMRFE) has become one of the leading methods, but its performance can be reduced because of the small sample size, noisy data and the fact that the method does not remove redundant genes.

Methods

We propose a novel framework for gene selection which uses the advantageous features of conventional methods and addresses their weaknesses. In fact, we have combined the Fisher method and SVMRFE to utilize the advantages of a filtering method as well as an embedded method. Furthermore, we have added a redundancy reduction stage to address the weakness of the Fisher method and SVMRFE. In addition to gene expression values, the proposed method uses Gene Ontology which is a reliable source of information on genes. The use of Gene Ontology can compensate, in part, for the limitations of microarrays, such as having a small number of samples and erroneous measurement results.

Results

The proposed method has been applied to colon, Diffuse Large B-Cell Lymphoma (DLBCL) and prostate cancer datasets. The empirical results show that our method has improved classification performance in terms of accuracy, sensitivity and specificity. In addition, the study of the molecular function of selected genes strengthened the hypothesis that these genes are involved in the process of cancer growth.

Conclusions

The proposed method addresses the weakness of conventional methods by adding a redundancy reduction stage and utilizing Gene Ontology information. It predicts marker genes for colon, DLBCL and prostate cancer with a high accuracy. The predictions made in this study can serve as a list of candidates for subsequent wet-lab verification and might help in the search for a cure for cancers.     

Identification of isopentenol biosynthetic genes from Bacillus subtilis by a screening method based on isoprenoid precursor toxicity

We have developed a novel method to clone terpene synthase genes. This method relies on the inherent toxicity of the prenyl diphosphate precursors to terpenes, which resulted in a reduced-growth phenotype. When these precursors were consumed by a terpene synthase, normal growth was restored. We have demonstrated that this method is capable of enriching a population of engineered Escherichia coli for those clones that express the sesquiterpene-producing amorphadiene synthase. In addition, we enriched a library of genomic DNA from the isoprene-producing bacterium Bacillus subtilis strain 6,051 in E. coli engineered to produce elevated levels of isopentenyl diphosphate and dimethylallyl diphosphate. The selection resulted in the discovery of two genes (yhfR and nudF) whose protein products acted directly on the prenyl diphosphate precursors and produced isopentenol. Expression of nudF in E. coli engineered with the mevalonate-based isopentenyl pyrophosphate biosynthetic pathway resulted in the production of isopentenol.