Alexander Ropelewski

Distributing the comparison of DNA and protein sequences across heterogeneous supercomputers

No abstract available

MPI-PHYLIP: Parallelizing Computationally Intensive Phylogenetic Analysis Routines for the Analysis of Large Protein Families

Background Phylogenetic study of protein sequences provides unique and valuable insights into the molecular and genetic basis of important medical and epidemiological problems as well as insights about the origins and development of physiological features in present day organisms. Consensus phylogenies based on the bootstrap and other resampling methods play a crucial part in analyzing the robustness of the trees produced for these analyses. Methodology Our focus was to increase the number of bootstrap replications that can be performed on large protein datasets using the maximum parsimony, distance matrix, and maximum likelihood methods. We have modified the PHYLIP package using MPI to enable large-scale phylogenetic study of protein sequences, using a statistically robust number of bootstrapped datasets, to be performed in a moderate amount of time. This paper discusses the methodology used to parallelize the PHYLIP programs and reports the performance of the parallel PHYLIP programs that are relevant to the study of protein evolution on several protein datasets. Conclusions Calculations that currently take a few days on a state of the art desktop workstation are reduced to calculations that can be performed over lunchtime on a modern parallel computer. Of the three protein methods tested, the maximum likelihood method scales the best, followed by the distance method, and then the maximum parsimony method. However, the maximum likelihood method requires significant memory resources, which limits its application to more moderately sized protein datasets.

Mathematically Complete Nucleotide and Protein Sequence Searching Using Ssearch

In this unit a protocol is described for predicting the structure of simple transmembrane a‐helical bundles. The protocol is based on a global molecular dynamics search (GMDS) of the configuration space of the helical bundle, yielding several candidate structures. The correct structure among these candidates is selected using information from silent amino acid substitutions, employing the premise that only the correct structure must (by definition) accept all of the silent amino acid substitutions. Thus, the correct structure is found by repeating the GMDS for several close homologs and selecting the structure that persists in all of the trials.

A mechanism for evolving novel plant sesquiterpene synthase function

Plant sesquiterpene synthases, a subset of the terpene synthase superfamily, are a mechanistically diverse family of enzymes capable of synthesizing hundreds of complex compounds with high regio‐ and stereospecificity and are of biological importance due to their role in plant defense mechanisms. In the current report we describe a large‐scale, high‐resolution phylogenetic analysis of ∼200 plant sesquiterpene synthases integrated with structural and experimental data that address these issues. We observe that all sequences that cluster together on the phylogenetic tree into well‐defined groups share at least the first reaction in the catalytic mechanism subsequent to the initial ionization step and many share steps beyond this down to proton transfers between the enzyme and substrate. Most significant is the previously unreported high conservation of an Asp‐Tyr‐Asp triad. Due to its high conservation, patterns in the phylogenetic tree as well as experimental and modeling results, we suggest that this Asp‐Tyr‐Asp triad is an important functional element responsible for many proton transfers to and from the substrate and intermediates along the plant sesquiterpene synthase catalytic cycle and whose position can be tuned by residues outside the active site that can lead to the evolution of novel enzyme function.

Implementation of Genetic Sequence Alignment Programs on Supercomputers

We have implemented exhaustive genetic sequence alignment codes on a variety of high performance computers. In this article, we compare and contrast the implementation issues encountered on different high performance computer architecture and the approaches used to overcome these problems. In addition, we discuss advanced sequence alignment techniques, including context sensitive and multiple sequence alignments.

De Novo Assembly of Lucina pectinata Genome using Ion Torrent Reads

Lucina pectinata is a bivalve that lives in sulfide-rich environments and houses intracellular sulfide oxidizing endosymbiont. This organism is an ideal model to understand adaptive mechanisms and chemoautotrophic endosymbiosis in organisms living in sulfide-rich environments. However, only three hemoglobins have been completely characterized at protein and gene level leaving a gap in understanding the biology of this organism. In this work, we produced draft genomic assemblies with data produced by the Ion Proton Next Generation Sequencing System using both the MIRA4 and SPAdes assemblers. We compare and contrast these draft assemblies using metrics such as N50, total assembled length, number of predicted genes and other measures. We conclude that de novo assembly of eukaryotic organisms with NGS data from the Ion technology family remains complicated and may benefit from the use of multiple genome assemblers.

In silico analysis of heme oxygenase structural homologues identifies group‐specific conservations

Heme oxygenases (HO) catalyze the breakdown of heme, aiding the recycling of its components. Several other enzymes have homologous tertiary structures to HOs, while sharing little sequence homology. These homologues include thiaminases, the hydroxylase component of methane monooxygenases, and the R2 component of Class I ribonucleotide reductases (RNR). This study compared these structural homologues of HO, using a large number of protein sequences for each homologue. Alignment of a total of 472 sequences showed little sequence conservation, with no residues having conservation in more than 80% of aligned sequences and only five residues conserved in at least 60% of the sequences. Fourteen additional positions, most of which were critical for hydrophobic packing, displayed amino acid similarity of 60% or higher. Ten conserved sequence motifs were identified in HOs and RNRs. Phylogenetic analysis verified the existence of the four distinct groups of HO homologues, which were then analyzed by group entropy analysis to identify residues critical to the unique function of each enzyme. Other methods for determining functional residues were also performed. Several common index positions identified represent critical evolutionary changes that resulted in the unique function of each enzyme, suggesting potential targets for site‐directed mutagenesis. These positions included residues that coordinate ligands, form the active sites, and maintain enzyme structure.

Assisting Bioinformatics Programs at Minority Institutions: Needs Assessment, and Lessons Learned -- A Look at an Internship Program

Over the last two decades, there has been a general acknowledgement within the scientific community that modern biology is becoming increasingly computational and to be prepared biologists need to build computational skills. We present work in assisting Bioinformatics efforts at minority institutions in the USA funded through a National Institutes of Health (NIH) grant over the last 15 years. The primary aim was to create a program for assisting minority institutions in building multidisciplinary bioinformatics training programs. The program involves four components for immediate and long-term increases in research opportunities at minority institutions. Specifically, we describe the results of a two month internship program. Through pre and post surveys reported by the participants, we have measured the skills levels of the internship participants prior to the training beginning and at the end of the training. The interns in the program have a stated interest in bioinformatics and are drawn exclusively from multiple minority serving institutions (MSIs) across the United States. The results of the incoming surveys indicate that the participants have acquired basic bioinformatics knowledge, but have not acquired general computational science skills needed to be successful practitioners within the field. This program has been a highly successful outreach effort and a very sound and cost-effective use of the Minority Access to Research Careers (MARC) funding program from NIH. Important lessons have been learned about bioinformatics education that should be implemented at the policy level in order to ensure that educators, students and researchers at minority serving institutions can address science problems using state-of-the-art computational methods, computational genomics and Big Data. We offer suggestions based on our experience in working with MSIs and with High Performance Computing (HPC) to help improve the preparation of students for careers as bioinformatics scientists.

A simple approach for the distribution of computationally intense tasks in an heterogeneous environment: distribution of the MDPP image-processing package

A method has been developed to link the display ability of a high-resolution graphics workstation with the computational power of a local mainframe or a remote supercomputer via an electronic data network. The method allows this link to be established in a manner largely transparent to the user. The application of the method is illustrated by our successful distribution of the computationally intensive portions of an imaging program (MDPP) from a small VAX workstation to a VAX mainframe and Cray Y-MP8/832 using a simple message-passing technique. This technique can be applied to almost any configuration of networked machines.

Applying iRODS to the Brain Image Library

Recent advancements in optical microscopy and specimen preparation have greatly improved the specificity and sensitivity of observation. Imaging instruments have become readily available and produce data at GB/s rates. These capabilities now allow capture of large numbers of whole-brain volumetric datasets, enabling morphological study including connectivity within the brain. Brain image datasets pose a problem for data storage, access, and analysis due to their large and complex structure. The Brain Image Library (BIL) is a national public resource of brain image datasets contributed by the brain research community. BIL is located at the Pittsburgh Supercomputing Center (PSC). To manage metadata associated with brain image collections, BIL uses iRODS, an open source data management system. This paper describes the use of iRODS for data management within the Brain Image Library an NIH BRAIN Initiative-funded petascale archive of spatial brain datasets.