In the era of data-driven science, conducting computational experiments that involve analysing large datasets using heterogeneous computational clusters, is part of the everyday routine for many scientists. Moreover, to ensure the credibility of their results, it is very important for these analyses to be easily reproducible by other researchers. Although various technologies, that could facilitate the work of scientists in this direction, have been introduced in the recent years, there is still a lack of open-source platforms that combine them to this end. In this work, we describe and demonstrate SCHeMa, an open-source platform that facilitates the execution and reproducibility of computational analysis on heterogeneous clusters, leveraging containerization, experiment packaging, workflow management, and machine learning technologies.

Abstract Motivation: Experimental evidence has accumulated showing that microRNA (miRNA) binding sites within protein coding sequences (CDSs) are functional in controlling gene expression. Results: Here we report a computational analysis of such miRNA target sites, based on features extracted from existing mammalian high-throughput immunoprecipitation and sequencing data. The analysis is performed independently for the CDS and the 3′-untranslated regions (3′-UTRs) and reveals different sets of features and models for the two regions. The two models are combined into a novel computational model for miRNA target genes, DIANA-microT-CDS, which achieves higher sensitivity compared with other popular programs and the model that uses only the 3′-UTR target sites. Further analysis indicates that genes with shorter 3′-UTRs are preferentially targeted in the CDS, suggesting that evolutionary selection might favor additional sites on the CDS in cases where there is restricted space on the 3′-UTR. Availability: The results of DIANA-microT-CDS are available at www.microrna.gr/microT-CDS Contact: [email protected]; [email protected] Supplementary information: Supplementary data are available at Bioinformatics online.

The in vivo activity of CA1 pyramidal neurons alternates between regular spiking and bursting, but how these changes affect information processing remains unclear. Using a detailed CA1 pyramidal neuron model, we investigate how timing and spatial arrangement variations in synaptic inputs to the distal and proximal dendritic layers influence the information content of model responses. We find that the temporal delay between activation of the two layers acts as a switch between excitability modes: short delays induce bursting while long delays decrease firing. For long delays, the average firing frequency of the model response discriminates spatially clustered from diffused inputs to the distal dendritic tree. For short delays, the onset latency and inter-spike-interval succession of model responses can accurately classify input signals as temporally close or distant and spatially clustered or diffused across different stimulation protocols. These findings suggest that a CA1 pyramidal neuron may be capable of encoding and transmitting presynaptic spatiotemporal information about the activity of the entorhinal cortex-hippocampal network to higher brain regions via the selective use of either a temporal or a rate code.

UNLABELLED: MicroRNAs (miRNAs) are a class of short endogenously expressed RNA molecules that regulate gene expression by binding directly to the messenger RNA of protein coding genes. They have been found to confer a novel layer of genetic regulation in a wide range of biological processes. Computational miRNA target prediction remains one of the key means used to decipher the role of miRNAs in development and disease. Here we introduce the basic idea behind the experimental identification of miRNA targets and present some of the most widely used computational miRNA target identification programs. The review includes an assessment of the prediction quality of these programs and their combinations. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

SUMMARY: DIANA-mirPath is a web-based computational tool developed to identify molecular pathways potentially altered by the expression of single or multiple microRNAs. The software performs an enrichment analysis of multiple microRNA target genes comparing each set of microRNA targets to all known KEGG pathways. The combinatorial effect of co-expressed microRNAs in the modulation of a given pathway is taken into account by the simultaneous analysis of multiple microRNAs. The graphical output of the program provides an overview of the parts of the pathway modulated by microRNAs, facilitating the interpretation and presentation of the analysis results. AVAILABILITY: The software is available at http://microrna.gr/mirpath and is free for all users with no login or download requirement.

BACKGROUND: MicroRNAs are small endogenously expressed non-coding RNA molecules that regulate target gene expression through translation repression or messenger RNA degradation. MicroRNA regulation is performed through pairing of the microRNA to sites in the messenger RNA of protein coding genes. Since experimental identification of miRNA target genes poses difficulties, computational microRNA target prediction is one of the key means in deciphering the role of microRNAs in development and disease. RESULTS: DIANA-microT 3.0 is an algorithm for microRNA target prediction which is based on several parameters calculated individually for each microRNA and combines conserved and non-conserved microRNA recognition elements into a final prediction score, which correlates with protein production fold change. Specifically, for each predicted interaction the program reports a signal to noise ratio and a precision score which can be used as an indication of the false positive rate of the prediction. CONCLUSION: Recently, several computational target prediction programs were benchmarked based on a set of microRNA target genes identified by the pSILAC method. In this assessment DIANA-microT 3.0 was found to achieve the highest precision among the most widely used microRNA target prediction programs reaching approximately 66%. The DIANA-microT 3.0 prediction results are available online in a user friendly web server at http://www.microrna.gr/microT.

For almost three decades, cancer was thought to result from changes in the structure and/or expression of protein coding genes. The discovery of thousands of genes that produce noncoding RNA (ncRNA) transcripts in the past few years suggested that the molecular biology of cancer is much more complex. MicroRNAs (miRNAs), an important group of ncRNAs, have recently been associated with tumorigenesis by acting either as tumor suppressors or oncogenes. Experimental prediction of miRNA genes is a slow process, because of the difficulties of cloning ncRNAs. Complementary to experimental approaches, a number of computational tools trained to recognize features of the biogenesis of miRNAs have significantly aided in the prediction of new miRNA candidates. By narrowing down the search space, computational approaches provide valuable clues as to which are the dominant features that characterize these regulatory units and which genes are their most likely targets. Moreover, through the use of high-throughput expression profiling methods, many molecular signatures of miRNA deregulation in human tumors have emerged. In this review, we present an overview of existing computational methods for identifying miRNA genes and assessing their expression levels, and analyze the contribution of such tools toward illuminating the role of miRNAs in cancer.

Cross-feeding, where two or more strains of an organism coexist on a single limiting resource, has been observed to emerge in long-term evolution experiments of E. coli in continuous culture. Here, we describe a computational method to model and systematically identify synergistic strains that have superior growth by exploiting their metabolic by-products with mutual benefit. A mutual flux balance simulation considers all possible single gene deletions growing on various substrates. Several synergistic strains are found to have higher growth than any single-strain cultures given the same limiting substrates. As the method is based on a detailed genome scale metabolic flux balance model of the organism, the results are not only consistent with several observed cross-feeding E. coli strains, but can also explain the exact mechanism of the synergy. We expect a broad range of applications for this method in metabolic engineering.

An algorithm called Bidirectional Long Short-Term Memory Networks (BLSTM) for processing sequential data is introduced. This supervised learning method trains a special recurrent neural network to use very long ranged symmetric sequence context using a combination of nonlinear processing elements and linear feedback loops for storing long-range context. The algorithm is applied to the sequence-based prediction of protein localization and predicts 93.3 percent novel non-plant proteins and 88.4 percent novel plant proteins correctly, which is an improvement over feedforward and standard recurrent networks solving the same problem. The BLSTM system is available as a web-service (http://www.stepc.gr/~synaptic/blstm.html).

MOTIVATION: Expressed Sequence Tags (ESTs) are next to cDNA sequences as the most direct way to locate in silico the genes of the genome and determine their structure. Currently ESTs make up more than 60% of all the database entries. The goal of this work is the development of a new program called DNA Intelligent Analysis for ESTs (DIANA-EST) based on a combination of Artificial Neural Networks (ANN) and statistics for the characterization of the coding regions within ESTs and the reconstruction of the encoded protein. RESULTS: 89.7% of the nucleotides from an independent test set with 127 ESTs were predicted correctly as to whether they are coding or non coding. AVAILABILITY: The program is available upon request from the author. CONTACT: Present address: Department of Genetics, University of Pennsylvania, School of Medicine, 475 Clinical Research Building, 415 Curie Boulevard, Philadelphia, PA 19104-6145, USA. [email protected].

This paper concerns a novel application of neural networks to magnetic resonance imaging (MRI) by considering regularized neural network models for the problem of image reconstruction from sparsely sampled k-space. Effective solutions to this problem are indispensable especially when dealing with MRI of dynamic phenomena since then, rapid sampling in k-space is required. The goal in such a case is to reduce the measurement time by omitting as many scanning trajectories as possible. This approach, however, entails underdetermined equations and leads to poor image reconstruction. It is proposed here that significant improvements could be achieved concerning image reconstruction if a procedure, based on neural network function approximation methodology and involving regularization techniques, for estimating the missing samples of complex k-space were introduced. To this end, the viability of involving neural network algorithms with/without regularization for such a problem is considered and it is found that their image reconstruction results are very favorably compared to the ones obtained by the trivial zero-filled k-space approach or traditional more sophisticated interpolation approaches. Moreover, it is found that regularized multilayer perceptrons outperform the ones not involving regularization during their training.

Expressed sequence tags (EST) are small parts of DNA, which are used to clone new genes. One main characteristic of EST is that they contain more than 1% sequencing errors. What we need to know is which parts of the EST contain information about proteins, the so called coding regions. In this paper we describe an error-tolerant program for the prediction of such coding regions in EST. The program is based on a combination of statistical methods and several artificial neural networks (ANN). 89.7% of the nucleotides of a independent test set with 127 EST's are predicted correctly as to whether they are coding or noncoding. These results are independent of the existence of homologous gene or protein sequences and representative for the application to the largest part of all EST.

We here present a parallel implementation of artificial neural networks on the connection machine CM-5 and compare it with other parallel implementations on SIMD and MIMD architectures. This parallel implementation was developed with the goal of efficiently training large neural networks with huge training pattern sets for applications in molecular biology, in particular the prediction of coding regions in DNA sequences. The implementation uses training pattern parallelism and makes use of the parallel I/O facilities of the CM-5 and its efficient reduction operations available within the control network to achieve a high scalability. The parallel simulator obtains a maximum speed of 149.25 MCUPS for training feedforward networks with backpropagation on a 512 processor CM-5 system without using the CM-5 vector facility. The implementation poses no restriction on the type of network topology and works with different batch training algorithms like BP. Quickprop and Rprop.