MOTIVATION: Protein-protein interaction (PPI), as a relative property, is determined by two binding proteins, which brings a great challenge to design an expert model with an unbiased learning architecture and a superior generalization performance. Additionally, few efforts have been made to allow PPI predictors to discriminate between relative properties and intrinsic properties. RESULTS: We present a sequence-based approach, DeepTrio, for PPI prediction using mask multiple parallel convolutional neural networks. Experimental evaluations show that DeepTrio achieves a better performance over several state-of-the-art methods in terms of various quality metrics. Besides, DeepTrio is extended to provide additional insights into the contribution of each input neuron to the prediction results. AVAILABILITY AND IMPLEMENTATION: We provide an online application at http://bis.zju.edu.cn/deeptrio. The DeepTrio models and training data are deposited at https://github.com/huxiaoti/deeptrio.git. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Rice blast disease caused by the fungus Magnaporthe grisea (M. grisea) is one of the most serious diseases for the cultivated rice Oryza sativa (O. sativa). A key factor causing rice blast disease and defense might be protein-protein interactions (PPIs) between rice and fungus. In this research, we have developed a computational pipeline to predict PPIs between blast fungus and rice. After cross-prediction by interolog-based and domain-based method, we achieved 532 potential PPIs between 27 fungus proteins and 236 rice proteins. Accuracy of jackknife test, 10-fold cross-validation test and independent test for these PPIs were 90.43, 93.85 and 84.67%, respectively, by using support vector machine classification method. Meanwhile, the pathogenic genes of blast fungus were enriched in the predicted PPIs network when compared with 1000 random interaction networks. The rice regulatory network was downloaded and divided into 228 subnetworks with over six nodes, and the top seven subnetworks affected by blast fungus through PPIs were investigated. The results indicated that 34 upregulated and 12 downregulated master regulators in rice interacting with the fungus proteins in response to the infection of blast fungus. The common master regulators in rice in response to the infection of M. grisea, Xanthomonas oryzae pv.oryzae and rice stripe virus were analyzed. The ubiquitin proteasome pathway was the common pathway in rice regulated by these three pathogens, while apoptosis signaling pathway was induced by fungus and bacteria. In summary, the results in this article provide insight into the process of blast fungus infection.

Heterotrimeric G protein signaling cascades are one of the primary metazoan sensing mechanisms linking a cell to environment. However, the number of experimentally identified effectors of G protein in plant is limited. We have therefore studied which tools are best suited for predicting G protein effectors in rice. Here, we compared the predicting performance of four classifiers with eight different encoding schemes on the effectors of G proteins by using 10-fold cross-validation. Four methods were evaluated: random forest, naive Bayes, K-nearest neighbors and support vector machine. We applied these methods to experimentally identified effectors of G proteins and randomly selected non-effector proteins, and tested their sensitivity and specificity. The result showed that random forest classifier with composition of K-spaced amino acid pairs and composition of motif or domain (CKSAAP_PROSITE_200) combination method yielded the best performance, with accuracy and the Mathew's correlation coefficient reaching 74.62% and 0.49, respectively. We have developed G-Effector, an online predictor, which outperforms BLAST, PSI-BLAST and HMMER on predicting the effectors of G proteins. This provided valuable guidance for the researchers to select classifiers combined with different feature selection encoding schemes. We used G-Effector to screen the effectors of G protein in rice, and confirmed the candidate effectors by gene co-expression data. Interestingly, one of the top 15 candidates, which did not appear in the training data set, was validated in a previous research work. Therefore, the candidate effectors list in this article provides both a clue for researchers as to their function and a framework of validation for future experimental work. It is accessible at http://bioinformatics.fafu.edu.cn/geffector.

RNA structure plays a crucial role in gene maturation, regulation and function. Determining the form and frequency of RNA folds is essential for a better understanding of how RNA exerts its functions. Low-throughput studies have focused on RNA primary sequences and expression levels, but with an emphasis on relatively small numbers of transcripts. However, with the recent advent of high-throughput technologies, it is realistic to begin analyzing RNA secondary structures on a genome-wide scale. Here, we review genome-wide RNA secondary structure profiles as well as advances in computational structure predictions. We further discuss the novel characteristics of RNA secondary structure across messenger RNAs. Probing RNA secondary structure by high-throughput sequencing will enable us to build atlases of RNA secondary structures, an important step in helping us to understand the versatility of RNA functions in diverse cellular processes.

Modern biology has moved from a science of individual measurements to a science where data are collected on an industrial scale. Foremost, among the new tools for biochemistry are chip arrays which, in one operation, measure hundreds of thousands or even millions of DNA sequences or RNA transcripts. While this is impressive, increasingly sophisticated analysis tools have been required to convert gene array data into gene expression levels. Despite the assumption that noise levels are low, since the number of measurements for an individual gene is small, identifying which signals are affected by noise is a priority. High-density oligonucleotide array (HDONAs) from NCBI GEO shows that, even in the best Human GeneChips 1/4 percent of data are affected by spatial noise. Earlier designs are noisier and spatial defects may affect more than 25 percent of probes. BioConductor R code is available as supplementary material which can be found on the Computer Society Digital Library at http://doi.ieeecomputersociety.org/10.1109/TCBB.2008.108 and via http://bioinformatics.essex.ac.uk/users/wlangdon/TCBB-2007-11-0161.tar.gz.

RNAnet [3] http://bioinformatics.essex.ac.uk/users/wlangdon/rnanet/ allows the user to calculate correlations of gene expression, both between genes and between components within genes. We investigate all of Ensembl http://www.ensembl.org and find all the Homo Sapiens exons for which there are sufficient robust Affymetrix HG-U133 Plus 2 GeneChip probes. Calculating correlation between mRNA probe measurements for the same exon shows many exons whose components are consistently up regulated and down regulated. However we identify other Ensembl exons where sub-regions within them are self consistent but these transcript blocks are not well correlated with other blocks in the same exon. We suggest many current Ensembl exon definitions are incomplete. Secondly, having identified exon with substructure we use machine learning to try and identify patterns in the DNA sequence lying between blocks of high correlation which might yield biological or technological explanations. A Backus-Naur form (BNF) context-free grammar constrains strongly typed genetic programming (STGP) to evolve biological motifs in the form of regular expressions (RE) (e.g. TCTTT) which classify gene exons with potential alternative mRNA expression from those without. We show biological patterns can be data mined by a GP written in gawk and using egrep from NCBI's GEO http://www.ncbi.nlm.nih.gov/geo/ database. The automatically produced DNA motifs suggest that alternative polyadenylation is not responsible. (Full version in TR-09-02 [7].) Blocky exons can be found in http://bioinformatics.essex.ac.uk/users/wlangdon/tr-09-02.tar.gz

The reliable interpretation of Affymetrix GeneChip data is a multi-faceted problem. The interplay between biophysics, bioinformatics and mining of GeneChip surveys is leading to new insights into how best to analyse the data. Many of the molecular processes occurring on the surfaces of GeneChips result from the high surface density of probes. Interactions between neighbouring adjacent probes affect their rate and strength of hybridization to targets. Competing targets may hybridize to the same probe, and targets may partially bind to more than one probe. The formation of these partial hybrids results in a number of probes not reaching thermodynamic equilibrium during hybridization. Moreover, some targets fold up, or cross-hybridize to other targets. Furthermore, probes may fold and can undergo chemical saturation. There are also sequence-dependent differences in the rates of target desorption during the washing stage. Improvements in the mappings between probe sequence and biological databases are leading to more accurate gene expression profiles. Moreover, algorithms that combine the intensities of multiple probes into single measures of expression are increasingly dependent upon models of the hybridization processes occurring on GeneChips. The large repositories of GeneChip data can be searched for systematic effects across many experiments. This data mining has led to the discovery of a family of thousands of probes, which show correlated expression across thousands of GeneChip experiments. These probes contain runs of guanines, suggesting that G-quadruplexes are able to form on GeneChips. We discuss the impact of these structures on the interpretation of data from GeneChip experiments.

We present an overview of image-processing methods for Affymetrix GeneChips. All GeneChips are affected to some extent by spatially coherent defects and image processing has a number of potential impacts on the downstream analysis of GeneChip data. Fortunately, there are now a number of robust and accurate algorithms, which identify the most disabling defects. One group of algorithms concentrate on the transformation from the original hybridisation DAT image to the representative CEL file. Another set uses dedicated pattern recognition routines to detect different types of hybridisation defect in replicates. A third type exploits the information provided by public repositories of GeneChips (such as GEO). The use of these algorithms improves the sensitivity of GeneChips, and should be a prerequisite for studies in which there are only few probes per relevant biological signal, such as exon arrays and SNP chips.

BACKGROUND: Affymetrix GeneChips are a popular platform for performing whole-genome experiments on the transcriptome. There are a range of different calibration steps, and users are presented with choices of different background subtractions, normalisations and expression measures. We wished to establish which of the calibration steps resulted in the biggest uncertainty in the sets of genes reported to be differentially expressed. RESULTS: Our results indicate that the sets of genes identified as being most significantly differentially expressed, as estimated by the z-score of fold change, is relatively insensitive to the choice of background subtraction and normalisation. However, the contents of the gene list are most sensitive to the choice of expression measure. This is irrespective of whether the experiment uses a rat, mouse or human chip and whether the chip definition is made using probe mappings from Unigene, RefSeq, Entrez Gene or the original Affymetrix definitions. It is also irrespective of whether both Present and Absent, or just Present, Calls from the MAS5 algorithm are used to filter genelists, and this conclusion holds for genes of differing intensities. We also reach the same conclusion after assigning genes to be differentially expressed using t-statistics, although this approach results in a large amount of false positives in the sets of genes identified due to the small numbers of replicates typically used in microarray experiments. CONCLUSION: The major calibration uncertainty that biologists need to consider when analysing Affymetrix data is how their multiple probe values are condensed into one expression measure.

BACKGROUND: Affymetrix GeneChip technology enables the parallel observations of tens of thousands of genes. It is important that the probe set annotations are reliable so that biological inferences can be made about genes which undergo differential expression. Probe sets representing the same gene might be expected to show similar fold changes/z-scores, however this is in fact not the case. RESULTS: We have made a case study of the mouse Surf4, chosen because it is a gene that was reported to be represented by the same eight probe sets on the MOE430A array by both Affymetrix and Bioconductor in early 2004. Only five of the probe sets actually detect Surf4 transcripts. Two of the probe sets detect splice variants of Surf2. We have also studied the expression changes of the eight probe sets in a public-domain microarray experiment. The transcripts for Surf4 are correlated in time, and similarly the transcripts for Surf2 are also correlated in time. However, the transcripts for Surf4 and Surf2 are not correlated. This proof of principle shows that observations of expression can be used to confirm, or otherwise, annotation discrepancies. We have also investigated groups of probe sets on the RAE230A array that are assigned to the same LocusID, but which show large variances in differential expression in any one of three different experiments on rat. The probe set groups with high variances are found to represent cases of alternative splicing, use of alternative poly(A) signals, or incorrect annotations. CONCLUSION: Our results indicate that some probe sets should not be considered as unique measures of transcription, because the individual probes map to more than one transcript dependent upon the biological condition. Our results highlight the need for care when assessing whether groups of probe sets all measure the same transcript.

We present CATHEDRAL, an iterative protocol for determining the location of previously observed protein folds in novel multidomain protein structures. CATHEDRAL builds on the features of a fast secondary-structure-based method (using graph theory) to locate known folds within a multidomain context and a residue-based, double-dynamic programming algorithm, which is used to align members of the target fold groups against the query protein structure to identify the closest relative and assign domain boundaries. To increase the fidelity of the assignments, a support vector machine is used to provide an optimal scoring scheme. Once a domain is verified, it is excised, and the search protocol is repeated in an iterative fashion until all recognisable domains have been identified. We have performed an initial benchmark of CATHEDRAL against other publicly available structure comparison methods using a consensus dataset of domains derived from the CATH and SCOP domain classifications. CATHEDRAL shows superior performance in fold recognition and alignment accuracy when compared with many equivalent methods. If a novel multidomain structure contains a known fold, CATHEDRAL will locate it in 90% of cases, with <1% false positives. For nearly 80% of assigned domains in a manually validated test set, the boundaries were correctly delineated within a tolerance of ten residues. For the remaining cases, previously classified domains were very remotely related to the query chain so that embellishments to the core of the fold caused significant differences in domain sizes and manual refinement of the boundaries was necessary. To put this performance in context, a well-established sequence method based on hidden Markov models was only able to detect 65% of domains, with 33% of the subsequent boundaries assigned within ten residues. Since, on average, 50% of newly determined protein structures contain more than one domain unit, and typically 90% or more of these domains are already classified in CATH, CATHEDRAL will considerably facilitate the automation of protein structure classification.

This paper reports a graph-theoretic program, GRATH, that rapidly, and accurately, matches a novel structure against a library of domain structures to find the most similar ones. GRATH generates distributions of scores by comparing the novel domain against the different types of folds that have been classified previously in the CATH database of structural domains. GRATH uses a measure of similarity that details the geometric information, number of secondary structures and number of residues within secondary structures, that any two protein structures share. Although GRATH builds on well established approaches for secondary structure comparison, a novel scoring scheme has been introduced to allow ranking of any matches identified by the algorithm. More importantly, we have benchmarked the algorithm using a large dataset of 1702 non-redundant structures from the CATH database which have already been classified into fold groups, with manual validation. This has facilitated introduction of further constraints, optimization of parameters and identification of reliable thresholds for fold identification. Following these benchmarking trials, the correct fold can be identified with the top score with a frequency of 90%. It is identified within the ten most likely assignments with a frequency of 98%. GRATH has been implemented to use via a server (http://www.biochem.ucl.ac.uk/cgi-bin/cath/Grath.pl). GRATH's speed and accuracy means that it can be used as a reliable front-end filter for the more accurate, but computationally expensive, residue based structure comparison algorithm SSAP, currently used to classify domain structures in the CATH database. With an increasing number of structures being solved by the structural genomics initiatives, the GRATH server also provides an essential resource for determining whether newly determined structures are related to any known structures from which functional properties may be inferred.