Microbial gene identification using interpolated Markov models.

This paper describes a new system, GLIMMER, for finding genes in microbial genomes. In a series of tests on Haemophilus influenzae , Helicobacter pylori and other complete microbial genomes, this system has proven to be very accurate at locating virtually all the genes in these sequences, outperforming previous methods. A conservative estimate based on experiments on H.pylori and H. influenzae is that the system finds >97% of all genes. GLIMMER uses interpolated Markov models (IMMs) as a framework for capturing dependencies between nearby nucleotides in a DNA sequence. An IMM-based method makes predictions based on a variable context; i.e., a variable-length oligomer in a DNA sequence. The context used by GLIMMER changes depending on the local composition of the sequence. As a result, GLIMMER is more flexible and more powerful than fixed-order Markov methods, which have previously been the primary content-based technique for finding genes in microbial DNA.     

<Emphasis Type="Italic">Ab initio</Emphasis> gene identification: Prokaryote genome annotation with GeneScan and GLIMMER

We compare the annotation of three complete genomes using theab initio methods of gene identification GeneScan and GLIMMER. The annotation given in GenBank, the standard against which these are compared, has been made using GeneMark. We find a number of novel genes which are predicted by both methods used here, as well as a number of genes that are predicted by GeneMark, but are not identified by either of the nonconsensus methods that we have used. The three organisms studied here are all prokaryotic species with fairly compact genomes. The Fourier measure forms the basis for an efficient non-consensus method for gene prediction, and the algorithm GeneScan exploits this measure. We have bench-marked this program as well as GLIMMER using 3 complete prokaryotic genomes. An effort has also been made to study the limitations of these techniques for complete genome analysis. GeneScan and GLIMMER are of comparable accuracy insofar as gene-identification is concerned, with sensitivities and specificities typically greater than 0.9. The number of false predictions (both positive and negative) is higher for GeneScan as compared to GLIMMER, but in a significant number of cases, similar results are provided by the two techniques. This suggests that there could be some as-yet unidentified additional genes in these three genomes, and also that some of the putative identifications made hitherto might require re-evaluation. All these cases are discussed in detail.     

Accuracy improvement for identifying translation initiation sites in microbial genomes

MOTIVATION: At present the computational gene identification methods in microbial genomes have a high prediction accuracy of verified translation termination site (3' end), but a much lower accuracy of the translation initiation site (TIS, 5' end). The latter is important to the analysis and the understanding of the putative protein of a gene and the regulatory machinery of the translation. Improving the accuracy of prediction of TIS is one of the remaining open problems. RESULTS: In this paper, we develop a four-component statistical model to describe the TIS of prokaryotic genes. The model incorporates several features with biological meanings, including the correlation between translation termination site and TIS of genes, the sequence content around the start codon; the sequence content of the consensus signal related to ribosomal binding sites (RBSs), and the correlation between TIS and the upstream consensus signal. An entirely non-supervised training system is constructed, which takes as input a set of annotated coding open reading frames (ORFs) by any gene finder, and gives as output a set of organism-specific parameters (without any prior knowledge or empirical constants and formulas). The novel algorithm is tested on a set of reliable datasets of genes from Escherichia coli and Bacillus subtillis. MED-Start may correctly predict 95.4% of the start sites of 195 experimentally confirmed E.coli genes, 96.6% of 58 reliable B.subtillis genes. Moreover, the test results indicate that the algorithm gives higher accuracy for more reliable datasets, and is robust to the variation of gene length. MED-Start may be used as a postprocessor for a gene finder. After processing by our program, the improvement of gene start prediction of gene finder system is remarkable, e.g. the accuracy of TIS predicted by MED 1.0 increases from 61.7 to 91.5% for 854 E.coli verified genes, while that by GLIMMER 2.02 increases from 63.2 to 92.0% for the same dataset. These results show that our algorithm is one of the most accurate methods to identify TIS of prokaryotic genomes. AVAILABILITY: The program MED-Start can be accessed through the website of CTB at Peking University: http://ctb.pku.edu.cn/main/SheGroup/MED_Start.htm.     

Finding prokaryotic genes by the 'frame-by-frame' algorithm: targeting gene starts and overlapping genes

MOTIVATION: Tightly packed prokaryotic genes frequently overlap with each other. This feature, rarely seen in eukaryotic DNA, makes detection of translation initiation sites and, therefore, exact predictions of prokaryotic genes notoriously difficult. Improving the accuracy of precise gene prediction in prokaryotic genomic DNA remains an important open problem. RESULTS: A software program implementing a new algorithm utilizing a uniform Hidden Markov Model for prokaryotic gene prediction was developed. The algorithm analyzes a given DNA sequence in each of six possible global reading frames independently. Twelve complete prokaryotic genomes were analyzed using the new tool. The accuracy of gene finding, predicting locations of protein-coding ORFs, as well as the accuracy of precise gene prediction, and detecting the whole gene including translation initiation codon were assessed by comparison with existing annotation. It was shown that in terms of gene finding, the program performs at least as well as the previously developed tools, such as GeneMark and GLIMMER. In terms of precise gene prediction the new program was shown to be more accurate, by several percentage points, than earlier developed tools, such as GeneMark.hmm, ECOPARSE and ORPHEUS. The results of testing the program indicated the possibility of systematic bias in start codon annotation in several early sequenced prokaryotic genomes. AVAILABILITY: The new gene-finding program can be accessed through the Web site: http:@dixie.biology.gatech.edu/GeneMark/fbf.cgi CONTACT: mark@amber.gatech.edu.     

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.     

Use of simulated data sets to evaluate the fidelity of metagenomic processing methods

Metagenomics is a rapidly emerging field of research for studying microbial communities. To evaluate methods presently used to process metagenomic sequences, we constructed three simulated data sets of varying complexity by combining sequencing reads randomly selected from 113 isolate genomes. These data sets were designed to model real metagenomes in terms of complexity and phylogenetic composition. We assembled sampled reads using three commonly used genome assemblers (Phrap, Arachne and JAZZ), and predicted genes using two popular gene-finding pipelines (fgenesb and CRITICA/GLIMMER). The phylogenetic origins of the assembled contigs were predicted using one sequence similarity-based (blast hit distribution) and two sequence composition-based (PhyloPythia, oligonucleotide frequencies) binning methods. We explored the effects of the simulated community structure and method combinations on the fidelity of each processing step by comparison to the corresponding isolate genomes. The simulated data sets are available online to facilitate standardized benchmarking of tools for metagenomic analysis.     

Experiment specific expression patterns

The differential analysis of genes between microarrays from several experimental conditions or treatments routinely estimates which genes change significantly between groups. As genes are never regulated individually, observed behavior may be a consequence of changes in other genes. Existing approaches like co-expression analysis aim to resolve such patterns from a wide range of experiments. The knowledge of such a background set of experiments can be used to compute expected gene behavior based on known links. It is particularly interesting to detect previously unseen specific effects in other experiments. Here, a new method to spot genes deviating from expected behavior (PAttern DEviation SCOring--Padesco) is devised. It uses linear regression models learned from a background set to arrive at gene specific prediction accuracy distributions. For a given experiment, it is then decided whether each gene is predicted better or worse than expected. This provides a novel way to estimate the experiment specificity of each gene. We propose a validation procedure to estimate the detection of such specific candidates and show that these can be identified with an average accuracy of about 85%.     

Identification of esophageal cancer pathway deviation and construction of a diagnosis model using three kernel genes

The purpose of this study is to better understand the role of interleukin 35 (IL35) in esophageal carcinoma by comparing the mRNA level in Barrett's esophageal mucosa and in matched normal squamous mucosa and to understand how the diagnosis model works with two other genes: hepatocyte nuclear factor 1B (HNF1B) and cAMP responsive element binding protein 3-like 1 (CREB3L1). By comparing carcinoma tissue and normal tissue samples, we extracted all the differentially expressed mRNAs. The bioinformatics analysis resulted in the discovery of three prominent genes. Eventually, the three genes were utilized to train a deep-learning model. An additional wet experiment was conducted to validate the effect of IL35. All the differentially expressed genes were enriched into nine groups, each of which has specific biological functions. Given that the three significant genes HNF1B, CREB3L1, and IL35 as diagnostic features, a deep-learning model was constructed, reaching an accuracy of 93% in the training set and 87% in the test set. Our findings suggest that IL35, along with the other two signatures, can distinguish esophageal tumor samples from normal samples precisely.     

Unsupervised feature selection algorithm for multiclass cancer classification of gene expression RNA-Seq data

This paper presents a Grouping Genetic Algorithm (GGA) to solve a maximally diverse grouping problem. It has been applied for the classification of an unbalanced database of 801 samples of gene expression RNA-Seq data in 5 types of cancer. The samples are composed by 20,531 genes. GGA extracts several groups of genes that achieve high accuracy in multiple classification. Accuracy has been evaluated by an Extreme Learning Machine algorithm and was found to be slightly higher in balanced databases than in unbalanced ones. The final classification decision has been made through a weighted majority vote system between the groups of features. The proposed algorithm finally selects 49 genes to classify samples with an average accuracy of 98.81% and a standard deviation of 0.0174.     

Recognition of 3' -processing sites of human mRNA precursors

We have developed a computer program POLYAH and an algorithmfor the identification of 3'-processing sites of human mRNAprecursors. The algorithm is based on a linear discriminantfunction (LDF) trained to discriminate real poly(A) signal regionsfrom the other regions of human genes possessing the AATAAAsequence which is most likely nonfunctional. As the parametersof LDF, various significant contextual characteristics of sequencessurrounding AATAAA signals were used. An accuracy of methodhas been estimated on a set of 131 poly(A) regions and 1466regions of human genes having the AATAAA sequence. When thethreshold was set to predict 86% of poly(A) regions correctly,specificity of 51% and correlation coefficient of 0.62 had beenachieved. The precision of this approach is better than forthe other methods and has been tested on a larger data set.POLYAH can be used through World Wide Web (at Gene-Finder Homepage: URL http: //dot.imgen.bcm.tmc.edu: 9331/gene-finder/ gf.html)or by sending files with uncharacterized human sequences tothe University of Houston or Weizmann Institute of Science e-mailservers.     

Taxonomic Resolutions Based on 18S rRNA Genes: A Case Study of Subclass Copepoda

Biodiversity studies are commonly conducted using 18S rRNA genes. In this study, we compared the inter-species divergence of variable regions (V1–9) within the copepod 18S rRNA gene, and tested their taxonomic resolutions at different taxonomic levels. Our results indicate that the 18S rRNA gene is a good molecular marker for the study of copepod biodiversity, and our conclusions are as follows: 1) 18S rRNA genes are highly conserved intra-species (intra-species similarities are close to 100%); and could aid in species-level analyses, but with some limitations; 2) nearly-whole-length sequences and some partial regions (around V2, V4, and V9) of the 18S rRNA gene can be used to discriminate between samples at both the family and order levels (with a success rate of about 80%); 3) compared with other regions, V9 has a higher resolution at the genus level (with an identification success rate of about 80%); and 4) V7 is most divergent in length, and would be a good candidate marker for the phylogenetic study of Acartia species. This study also evaluated the correlation between similarity thresholds and the accuracy of using nuclear 18S rRNA genes for the classification of organisms in the subclass Copepoda. We suggest that sample identification accuracy should be considered when a molecular sequence divergence threshold is used for taxonomic identification, and that the lowest similarity threshold should be determined based on a pre-designated level of acceptable accuracy.     

Recognizing shorter coding regions of human genes based on the statistics of stop codons.

With the quick progress of the Human Genome Project, a great amount of uncharacterized DNA sequences needs to be annotated copiously by better algorithms. Recognizing shorter coding sequences of human genes is one of the most important problems in gene recognition, which is not yet completely solved. This paper is devoted to solving the issue using a new method. The distributions of the three stop codons, i.e., TAA, TAG and TGA, in three phases along coding, noncoding, and intergenic sequences are studied in detail. Using the obtained distributions and other coding measures, a new algorithm for the recognition of shorter coding sequences of human genes is developed. The accuracy of the algorithm is tested based on a larger database of human genes. It is found that the average accuracy achieved is as high as 92.1% for the sequences with length of 192 base pairs, which is confirmed by sixfold cross-validation tests. It is hoped that by incorporating the present method with some existing algorithms, the accuracy for identifying human genes from unannotated sequences would be increased.     

Hybrid Method Based on Information Gain and Support Vector Machine for Gene Selection in Cancer Classi?cation

It remains a great challenge to achieve suf?cient cancer classi?cation accuracy with the entire set of genes, due to the high dimensions, small sample size, and big noise of gene expression data. We thus proposed a hybrid gene selection method, Information Gain-Support Vector Machine(IG-SVM) in this study. IG was initially employed to ?lter irrelevant and redundant genes. Then,further removal of redundant genes was performed using SVM to eliminate the noise in the datasets more effectively. Finally, the informative genes selected by IG-SVM served as the input for the LIBSVM classi?er. Compared to other related algorithms, IG-SVM showed the highest classi?cation accuracy and superior performance as evaluated using ?ve cancer gene expression datasets based on a few selected genes. As an example, IG-SVM achieved a classi?cation accuracy of 90.32% for colon cancer, which is dif?cult to be accurately classi?ed, only based on three genes including CSRP1, MYL9, and GUCA2B.     

Recognition of translation initiation sites of eukaryotic genes based on an EM algorithm.

With the rapid increase of DNA databases of human and other eukaryotic model organisms, a large great number of genes need to be distinguished from the DNA databases. Exact recognition of translation initiation sites (TISs) of eukaryotic genes is very important to understand the translation initiation process, predict the detailed structure of eukaryotic genes, and annotate uncharacterized sequences. The problem has not been solved satisfactorily, especially for recognizing TISs of the eukaryotic genes with shorter first exons. It is an important task for extracting new features and finding new powerful algorithms for recognizing TISs of eukaryotic genes. In this paper, the important characteristics of shorter flanking fragments around TISs are extracted and an expectation-maximization (EM) algorithm based on incomplete data is used to recognize TISs of eukaryotic genes. The accuracy is up to 87.8% over a six-fold cross-validation test. The result shows that the identification variables are effectively extracted and the EM algorithm is a powerful tool to predict the TISs of eukaryotic genes. The algorithm also can be applied to other classification or clustering tasks in bioinformatics.     

The characterisation of three types of genes that overlie copy number variable regions

Background

Due to the increased accuracy of Copy Number Variable region (CNV) break point mapping, it is now possible to say with a reasonable degree of confidence whether a gene (i) falls entirely within a CNV; (ii) overlaps the CNV or (iii) actually contains the CNV. We classify these as type I, II and III CNV genes respectively.

Principal Findings

Here we show that although type I genes vary in copy number along with the CNV, most of these type I genes have the same expression levels as wild type copy numbers of the gene. These genes must, therefore, be under homeostatic dosage compensation control. Looking into possible mechanisms for the regulation of gene expression we found that type I genes have a significant paucity of genes regulated by miRNAs and are not significantly enriched for monoallelically expressed genes. Type III genes, on the other hand, have a significant excess of genes regulated by miRNAs and are enriched for genes that are monoallelically expressed.

Significance

Many diseases and genomic disorders are associated with CNVs so a better understanding of the different ways genes are associated with normal CNVs will help focus on candidate genes in genome wide association studies.     

An integrated cross-platform prognosis study on neuroblastoma patients

There have been several reports about the potential for predicting prognosis of neuroblastoma patients using microarray gene expression profiling of the tumors. However these studies have revealed an apparent diversity in the identity of the genes in their predictive signatures. To test the contribution of the platform to this discrepancy we applied the z-scoring method to minimize the impact of platform and combine gene expression profiles of neuroblastoma (NB) tumors from two different platforms, cDNA and Affymetrix. A total of 12442 genes were common to both cDNA and Affymetrix arrays in our data set. Two-way ANOVA analysis was applied to the combined data set for assessing the relative effect of prognosis and platform on gene expression. We found that 26.6% (3307) of the genes had significant impact on survival. There was no significant impact of microarray platform on expression after application of z-scoring standardization procedure. Artificial neural network (ANN) analysis of the combined data set in a leave-one-out prediction strategy correctly predicted the outcome for 90% of the samples. Hierarchical clustering analysis using the top-ranked 160 genes showed the great separation of two clusters, and the majority of matched samples from the different platforms were clustered next to each other. The ANN classifier trained with our combined cross-platform data for these 160 genes could predict the prognosis of 102 independent test samples with 71% accuracy. Furthermore it correctly predicted the outcome for 85/102 (83%) NB patients through the leave-one-out cross-validation approach. Our study showed that gene expression studies performed in different platforms could be integrated for prognosis analysis after removing variation resulting from different platforms.     

A probabilistic method for identifying start codons in bacterial genomes.

As the pace of genome sequencing has accelerated, the need for highly accurate gene prediction systems has grown. Computational systems for identifying genes in prokaryotic genomes have sensitivities of 98-99% or higher (Delcher et al., Nucleic Acids Res., 27, 4636-4641, 1999). These accuracy figures are calculated by comparing the locations of verified stop codons to the predictions. Determining the accuracy of start codon prediction is more problematic, however, due to the relatively small number of start sites that have been confirmed by independent, non-computational methods. Nonetheless, the accuracy of gene finders at predicting the exact gene boundaries at both the 5' and 3' ends of genes is of critical importance for microbial genome annotation, especially in light of the important signaling information that is sometimes found on the 5' end of a protein coding region. In this paper we propose a probabilistic method to improve the accuracy of gene identification systems at finding precise translation start sites. The new system, RBSfinder, is tested on a validated set of genes from Escherichia coli, for which it improves the accuracy of start site locations predicted by computational gene finding systems from the range 67-77% to 90% correct.     

Prioritization of differentially expressed genes through integrating public expression data

Differentially expressed gene (DEG) analysis is a major approach for interpreting phenotype differences and produces a large number of candidate genes. Given that it is burdensome to validate too many genes through benchwork, an urgent need exists for DEG prioritization. Here, a novel method is proposed for prioritizing bona fide DEGs by constructing the normal range of gene expression through integrating public expression data. Prioritization was performed by ranking the differences in cumulative probability for genes in case and control groups. DEGs from a study on pig muscle tissue were used to evaluate the prioritization accuracy. The results showed that the method reached an area under the receiver operating characteristic curve of 96.42% and can effectively shorten the list of candidate genes from a differential expression experiment to find novel causal genes. Our method can be easily extended to other tissues or species to promote functional research in broad applications.