BioPerf: A benchmark suite to evaluate high-performance computer architecture on bioinformatics applications

Adoption of an improved PSO to explore a compound multi-objective energy function in protein structure prediction

BackgroundProtein structure comparison is one of the most important problems in computational biology and plays a key role in protein structure prediction, fold family classification, motif finding, phylogenetic tree reconstruction and protein docking.ResultsWe propose a novel method to compare the protein structures in an accurate and efficient manner. Such a method can be used to not only reveal divergent evolution, but also identify circular permutations and further detect active-sites. Specifically, we define the structure alignment as a multi-objective optimization problem, i.e., maximizing the number of aligned atoms and minimizing their root mean square distance. By controlling a single distance-related parameter, theoretically we can obtain a variety of optimal alignments corresponding to different optimal matching patterns, i.e., from a large matching portion to a small matching portion. The number of variables in our algorithm increases with the number of atoms of protein pairs in almost a linear manner. In addition to solid theoretical background, numerical experiments demonstrated significant improvement of our approach over the existing methods in terms of quality and efficiency. In particular, we show that divergent evolution, circular permutations and active-sites (or structural motifs) can be identified by our method. The software SAMO is available upon request from the authors, or from and .ConclusionA novel formulation is proposed to accurately align protein structures in the framework of multi-objective optimization, based on a sequence order-independent strategy. A fast and accurate algorithm based on the bipartite matching algorithm is developed by exploiting the special features. Convergence of computation is shown in experiments and is also theoretically proven.

Protein structure prediction aims to model the three‐dimensional (3D) structure of so far structurally uncharacterised proteins from their amino acid sequence. Motivated by the observation that homologous proteins with related amino acid sequences have similar 3D structures, protein homology modelling uses comparative methods to generate models for a target protein based on one or more related proteins with known 3D structure. The coordinates of the model are generated based on alignments between the target's and template's amino acid sequences, which define the correspondence between residues in both proteins. Ultimately, the quality of a computational model determines its usefulness for specific biomedical applications. Therefore, model quality estimation methods are used to identify unreliable or erroneous regions in the resulting models, and to estimate the overall accuracy of a model. Homology modelling (or comparative modelling) is currently the most accurate computational method available to routinely generate models of sufficient quality for various applications in life science research. Comparative protein modelling methods have been completely automated in recent years, and several Internet servers offer protein modelling services which are reliable and easy to use – also for the nonexpert in computational biology. Key Concepts: Protein structure prediction aims to model the three‐dimensional structure of so far structurally uncharacterised proteins (‘target’) based on their amino acid sequence. Homologous proteins with related amino acid sequences have similar three‐dimensional structures. Protein homology modelling uses information from one or more related proteins with known three‐dimensional structure (‘template’) to generate models for the target protein. Sensitive sequence searching methods are applied to identify template proteins with known structures in large databases. An alignment between the target's and template's amino acid sequences describes the correspondence between residues in both proteins. The coordinates of the model are constructed by extracting positional information from the corresponding structural template. Segments of the target protein not covered by template information (e.g. insertions/deletion in the alignment) have to be constructed using de novo modelling methods. Model quality estimation methods are used to identify unreliable or erroneous regions in the resulting models. Ultimately, the quality of a structural model determines its usefulness for specific biomedical applications.

Proteins are essential parts of our life and participate in virtually every process within a cell. The understanding of protein structures is vital to determine the function of a protein. Protein structure prediction (PSP) from amino acid sequence is one of the high focus problems in bioinformatics today. This is due to the fact that the biological function of the protein is determined by its three dimensional structure. Thus, protein structure prediction is a fundamental area of computational biology. Its importance is intensed by large amounts of sequence data coming from PDB (Protein Data Bank) and the fact that experimentally methods such as X-ray crystallography or Nuclear Magnetic Resonance (NMR)which are used to determining protein structures remains very expensive and time consuming. For minimizing the time, computational methods are used for protein folding and structure prediction problem. In this paper results of protein p53 are discussed.

With a wealth of biodiversity, a long tradition of agriculture-based industries, and an established medical and biotechnological research and development community, Thailand has become an attractive location for life sciences investment. The large amount of data generated in many areas of life sciences requires visualization, management, and analysis, principally through bioinformatics. To become successful, Thailand's research community should emphasize establishing core technologies, such as genomics and bioinformatics, to boost development of agriculture, food processing, and biomedical research. The Thai government realized the importance of this field and created a national policy to greatly increase Thailand's participation in bioinformatics and genomics, budgeting for specific development goals in research infrastructure, education, and sustainable human resources. Thailand has not lagged behind in bioinformatics research activity and recognizes the importance of bioinformatics through increased policy awareness, human resources development, and increased research activity involving genomic-scale data generation and computational analyses. Many applications of genomics and bioinformatics to biomedical research and development in Thailand have progressed substantially during the past few years, leading to successful applications in some specific local areas. However, the applications to other important areas, such as agriculture, are hampered by the limited availability of genomic sequence data and the lack of necessary biochemical/physiological information. With the advent of more and more genomic information in public databases, Thailand's research community is striving to adopt comparative genomics to obtain information of direct relevance to the country's health and industrial needs. This article highlights Thailand's contribution to genomics and bioinformatics in the following areas: (1) policy support from the Thai government, (2) capacity building through infrastructure/education/human resources, and (3) research and development in genomics and computational biology. (See Box 1 for Authors' Biographies). Box 1. Authors' Biographies Wannipha Tongsima, M.S., obtained her master's degree in Industrial Microbiology from Chulalongkorn University, Thailand. She was involved in founding the Bioinformatics research program in BIOTEC. To reinforce the research activity in this area, she also helped organize the first International Conference on Computational Biology (InCoB), held in Bangkok in 2002. Later, she was appointed to manage one of the first BIOTEC ethnic-specific human genetic variation programs, named the Thailand SNP Discovery Project. She works as a Genomic Medicine program coordinator for the Cluster and Program Management Office (CPMO) of the National Science and Technology Development Agency (NSTDA), which is an umbrella organization of four other national research centers in Thailand, including BIOTEC. Sissades Tongsima, Ph.D., received his doctoral degree in Computer Science and Engineering from the University of Notre Dame, Indiana, United States. He has worked for the National Electronics and Computer Technology Center on High Performance Computing (HPC) and Computational Grid. During 2002–2004, he cochaired the Asia-Pacific Advanced Network (APAN) Grid Working Group. In 2003, he shifted his research direction from HPC architecture to bioinformatics research, when he started working for BIOTEC, and constructed the ThaiSNP database. His main research interest is in developing algorithms and databases for analyzing various research projects on human genetic variation. He currently heads the Genome Institute biostatistics and informatics laboratory at BIOTEC. Prasit Palittapongarnpim, M.D., earned his medical degree from Mahidol University, Thailand, and his B.S. in Mathematics from Ramkumhang University, Thailand. He is a Fellow of the Royal College of Pediatricians of Thailand and also an Associate Professor in Microbiology at Mahidol University, where he has conducted research focusing on tuberculosis. While holding a Deputy Director position, he initiated the Bioinformatics research program at BIOTEC in 2002 and led the organization of the first InCoB conference in 2002. He is currently a Vice President of NSTDA.

It is impossible to pinpoint the exact moment at which computational biology became a discipline of its own, but one could say that it was in 1997 when the society of computational biology was formed. Regardless of its exact birthday, the research community has rapidly adopted computational biology and its applications are being vigorously explored. The study and application of medicine is a dynamic challenge. Changes in medicine usually take place as a result of new knowledge acquired through observation and experimentation. When a tamping rod 1-inch thick went through Phineas Gage’s head in 1848, his survival gave the medical field an unusual opportunity to observe behavior of a person missing their prefrontal cortex. This observation lead to the short-lived psychosurgical procedure known as a lobotomy, which attempted to change a person’s behavior by separating two portions of a person’s brain (Pols, 2001). Countless observations, experiments and mistakes represent how almost all medical knowledge has been acquired. The relatively new field of computational biology offers a nontraditional approach to contribute to the medical body of knowledge. Computational biology is a new field combining biology, computer science, and mathematics to solve problems that are unworkable with traditional biological techniques. It includes traditional areas such as systems biology, molecular biology, biochemistry, biophysics, statistics, and computer science, as well as recently developed disciplines including bioinformatics and computational genomics. Algorithms, which are able to closely model biological behavior, validate the medical understanding of the observed processes and can be used to model scenarios that might not be able to be physically reproduced. The goal of computational biology is to use mathematics and computer science to model biological systems on the molecular level. Instead of taking on large complex systems, computational biology is starting small, literally. Modeling problems in molecular biology and biochemistry is a far less daunting task. At a microscopic level, patient’s characteristics drop out of the equation and all information behavior affecting is known. This creates a deterministic model which, given the same input, will always produce the same output. Some of the major subdisciplines of computational biology are computational genomics, systems biology, protein structure prediction, and evolutionary biology, all of which model microscopic structures.

Modern polyhedral compilers excel at aggressively optimizing codes with static control parts, but the state-of-practice to find high-performance polyhedral transformations especially for different hardware targets still largely involves auto-tuning. In this work we propose a novel customizable polyhedral scheduling technique, with the aim of delivering high performance for several hardware targets. We design constraints and objectives that model several crucial aspects of performance such as stride optimization or the trade-off between parallelism and reuse, while considering important architectural features of the target machine. We evaluate our work using the PolyBench/C benchmark suite and experimentally validate it against large optimization spaces generated with the Pluto compiler on 3 representative architectures: an IBM Power9, an Intel Xeon Phi and an Intel Core-i9. Our results show we can achieve comparable or superior performance to Pluto on the majority of benchmarks, without implementing tiling in the source code nor using experimental autotuning.

gem5 has been extensively used in computer architecture simulations and in the evaluation of new architectures for HPC (high performance computing) systems. Previous work has validated gem5 against ARM platforms. However, gem5 still shows high inaccuracy when modeling Intel’s x86 based processors. In this work, we focus on the simulation of a single node high performance system and study the sources of inaccuracies of gem5. Then we validate gem5 simulator against an Intel processor, Core-i7 (Haswell microarchitecture). We configured gem5 as close as possible to match Core-i7 Haswell microarchitecture configurations and made changes to the simulator to add some features, modified existing code, and tuned built-in configurations. As a result, we validated the simulator by fixing many sources of errors to match real hardware results with less than 6% mean error rate for different control, memory, dependency and execution microbenchmarks.

The NAS Parallel Benchmarks for evaluating C++ parallel programming frameworks on shared-memory architectures

Protein structure prediction is one of the most important problems in modern computational biology. Protein secondary structure prediction is a key step in prediction of protein tertiary structure. There have emerged many methods based on machine learning techniques, such as neural networks (NN) and support vector machine (SVM) etc., to focus on the prediction of the secondary structures. In this paper, a new method was proposed based on SVM. Different from the existing methods, this method takes into account of the physical-chemical properties and structure properties of amino acids. When tested on the most popular dataset CB513, it achieved a Q(3) accuracy of 0.7844, which illustrates that it is one of the top range methods for protein of secondary structure prediction.

Protein structure prediction is one of the most challenging problems in computational biology. Given a protein's amino acid sequence, a simplified version of the problem is to find an on-lattice self-avoiding walk that minimizes the interaction energy among the amino acids. In this paper, we present a memory-based local search method for the simplified problem using Hydrophobic-Polar energy model and Face Centered Cubic lattice. By memorizing local minima and then avoiding their neighbohood, our approach significantly improves the state-of-the-art local search method for protein structure prediction on a set of standard benchmark proteins.

AlphaFold2 (AF2) is an artificial intelligence (AI) system developed by DeepMind that can predict three-dimensional (3D) structures of proteins from amino acid sequences with atomic-level accuracy. Protein structure prediction is one of the most challenging problems in computational biology and chemistry, and has puzzled scientists for 50 years. The advent of AF2 presents an unprecedented progress in protein structure prediction and has attracted much attention. Subsequent release of structures of more than 200 million proteins predicted by AF2 further aroused great enthusiasm in the science community, especially in the fields of biology and medicine. AF2 is thought to have a significant impact on structural biology and research areas that need protein structure information, such as drug discovery, protein design, prediction of protein function, et al. Though the time is not long since AF2 was developed, there are already quite a few application studies of AF2 in the fields of biology and medicine, with many of them having preliminarily proved the potential of AF2. To better understand AF2 and promote its applications, we will in this article summarize the principle and system architecture of AF2 as well as the recipe of its success, and particularly focus on reviewing its applications in the fields of biology and medicine. Limitations of current AF2 prediction will also be discussed.

The prediction of a minimum energy protein structure from its amino acid sequence represents an important and challenging problem in computational biology. In this paper, we propose a novel heuristic approach for protein structure prediction (PSP) based on the concept of optimal hydrophobic core formation. Using 2D HP model, a well-known set of sub-structures analogous to the secondary structures are obtained. Some sub-conformations are appropriately classified and then incorporated as prior knowledge. Unlike most of the popular PSP approaches which are stochastic in nature, the proposed method is deterministic. The effectiveness of the proposed algorithm is evaluated by well-known benchmark as well as non-benchmark sequences commonly used with 2D HP model. Maintaining similar accuracy as other core based and population based algorithms our method is significantly faster and reduces the computation time as it avoids blind search within the hydrophobic core (H-Core).

AIMOES: Archive information assisted multi-objective evolutionary strategy for ab initio protein structure prediction

Human beings need some fundamental requirements to live, including food, drink, air, and shelter. Ever since the dawn of civilisation, they’ve been a source of contention among people. However, despite the rapid advancement of technology and the abundance of comforts available in today’s society, some basic needs have remained the same. This “fight for existence” continues to this day and will continue indefinitely. Pests and diseases often threaten the food crops lead to loss of production, availability of food which in turn affects the global economy. Modern molecular biology, microarrays, nanotechnology, and other advanced analytical techniques are helping to better understand the target pests and diseases. To cope with agricultural pests, as well as diseases that may be fatal, pest management techniques are being developed and modified in recent years. Current data suggest that further important plant protection tools would be introduced soon which could reduce the reliance on just insecticides to solve this issue. Computational biology, as an integrated subject, is increasingly finding its way into the development of pest control strategies. Fundamental data analysis in the research and development sector, as well as effective mass communication and information dissemination technologies, is used to establish direct connections with farmers and other grassroots people. Because of its enormous size, this stream has been found very essential and helpful in every respect. Bioinformatics is used in several pest control methods, either directly or indirectly, in this article few examples like pesticide design and screening, pesticide production optimisation, and pesticide toxicity prediction for a variety of chemicals have been discussed. We are moving in a new direction in pest control due to the current state of things as well as its popularity and applicability. In today’s climate, pest control requires more interdisciplinary contributions than ever before. As a consequence, the noble cause of food grain preservation before and after harvest should be addressed to feed a bigger population with less economic loss and to reduce sickness and high mortality rates. With the advent of data analytics, technology has become more prevalent in agriculture. One of the main challenges in using technology in agriculture is determining the effectiveness of big data analytics algorithms and their application methods.KeywordsBioinformaticsPestSoftwareDatabase