High-Accuracy ncRNA Function Prediction via Deep Learning Using Global and Local Sequence Information.

The prediction of the biological function of non-coding ribonucleic acid (ncRNA) is an important step towards understanding the regulatory mechanisms underlying many diseases. Since non-coding RNAs are present in great abundance in human cells and are functionally diverse, developing functional prediction tools is necessary. With recent advances in non-coding RNA biology and the availability of complete genome sequences for a large number of species, we now have a window of opportunity for studying non-coding RNA biology. However, the computational methods used to predict the non-coding RNA functions are mostly either scarcely accurate, when based on sequence information alone, or prohibitively expensive in terms of computational burden when a secondary structure prediction is needed. We propose a novel computational method to predict the biological function of non-coding RNA genes that is based on a collection of deep network architectures utilizing solely ncRNA sequence information and which does not rely on or require expensive secondary ncRNA structure information. The approach presented in this work exhibits comparable or superior accuracy to methods that employ both sequence and structural features, at a much lower computational cost.

A Drug Discovery Approach for an Effective Pain Therapy through Selective Inhibition of Nav1.7.

Chronic pain is a widespread disorder affecting millions of people and is insufficiently addressed by current classes of analgesics due to significant long-term or high dosage side effects. A promising approach that was recently proposed involves the systemic inhibition of the voltage-gated sodium channel Nav1.7, capable of cancelling pain perception completely. Notwithstanding numerous attempts, currently no drugs have been approved for the inhibition of Nav1.7. The task is complicated by the difficulty of creating a selective drug for Nav1.7, and avoiding binding to the many human paralogs performing fundamental physiological functions. In our work, we obtained a promising set of ligands with up to 5-40-fold selectivity and reaching 5.2 nanomolar binding affinity by employing a proper treatment of the problem and an innovative differential in silico screening procedure to discriminate for affinity and selectivity against the Nav paralogs. The absorption, distribution, metabolism, and excretion (ADME) properties of our top-scoring ligands were also evaluated, with good to excellent results. Additionally, our study revealed that the top-scoring ligand is a stereoisomer of an already-approved drug. These facts could reduce the time required to bring a new effective and selective Nav1.7 inhibitor to the market.

BGBlast: a BLAST grid implementation with database self-updating and adaptive replication.

BLAST is probably the most used application in bioinformatics teams. BLAST complexity tends to be a concern when the query sequence sets and reference databases are large. Here we present BGBlast: an approach for handling the computational complexity of large BLAST executions by porting BLAST to the Grid platform, leveraging the power of the thousands of CPUs which compose the EGEE infrastructure. BGBlast provides innovative features for efficiently managing BLAST databases in the distributed Grid environment. The system (1) keeps the databases constantly up to date while still allowing the user to regress to earlier versions, (2) stores the older versions of databases on the Grid with a time and space efficient delta encoding and (3) manages the number of replicas for each database over the Grid with an adaptive algorithm, dynamically balancing between execution parallelism and storage costs.

Data handling strategies for high throughput pyrosequencers.

BACKGROUND: New high throughput pyrosequencers such as the 454 Life Sciences GS 20 are capable of massively parallelizing DNA sequencing providing an unprecedented rate of output data as well as potentially reducing costs. However, these new pyrosequencers bear a different error profile and provide shorter reads than those of a more traditional Sanger sequencer. These facts pose new challenges regarding how the data are handled and analyzed, in addition, the steep increase in the sequencers throughput calls for much computation power at a low cost. RESULTS: To address these challenges, we created an automated multi-step computation pipeline integrated with a database storage system. This allowed us to store, handle, index and search (1) the output data from the GS20 sequencer (2) analysis projects, possibly multiple on every dataset (3) final results of analysis computations (4) intermediate results of computations (these allow hand-made comparisons and hence further searches by the biologists). Repeatability of computations was also a requirement. In order to access the needed computation power, we ported the pipeline to the European Grid: a large community of clusters, load balanced as a whole. In order to better achieve this Grid port we created Vnas: an innovative Grid job submission, virtual sandbox manager and job callback framework. After some runs of the pipeline aimed at tuning the parameters and thresholds for optimal results, we successfully analyzed 273 sequenced amplicons from a cancerous human sample and correctly found punctual mutations confirmed by either Sanger resequencing or NCBI dbSNP. The sequencing was performed with our 454 Life Sciences GS 20 pyrosequencer. CONCLUSION: We handled the steep increase in throughput from the new pyrosequencer by building an automated computation pipeline associated with database storage, and by leveraging the computing power of the European Grid. The Grid platform offers a very cost effective choice for uneven workloads, typical in many scientific research fields, provided its peculiarities can be accepted (these are discussed). The mentioned infrastructure was used to analyze human amplicons for mutations. More analyses will be performed in the future.