Carlos R. Erig Lima

A CONFIGWARE APPROACH FOR HIGH-SPEED PARALLEL ANALYSIS OF GENOMIC DATA

Many problems in bioinformatics represent great computational challenges due to the huge amount of biological data to be analyzed. Reconfigurable systems can offer custom-computing machines, with orders of magnitude faster than regular software, running in general-purpose processors. We present a methodology for using a configware system in an interesting problem of molecular biology: the splice junction detection in eukaryote genes. Decision trees were developed using a benchmark of DNA sequences. They were converted into logical equations, simplified, and submitted to a Boolean minimization. The resulting circuit was implemented in reconfigurable parallel hardware and evaluated with a five-fold cross-validation procedure, run in a second level of parallelism. The average accuracy achieved was 90.41% and it takes 18 ns to process each data record with 60 nucleotides.

Reconfigurable Computing for Accelerating Protein Folding Simulations

This paper presents a methodology for the design of a reconfigurable computing system applied to a complex problem in molecular Biology: the protein folding problem. An efficient hardware-based approach was devised to achieve a significant reduction of the search space of possible foldings. Several simulations were done to evaluate the performance of the system as well as the demand for FPGA’s resources. Also a complete desktop-FPGA system was developed. This work is the base for future hardware implementations aimed at finding the optimal solution for protein folding problems using the 2D-HP model.

A comparison of FPGA and FPAA technologies for a signal processing application

This paper presents a comparison between two technologies for reconfigurable circuits: FPGAs and FPAAs. The comparison is based on a case study of the area of industrial control using simulations with both types of reconfigurable devices. Several design issues are discussed, including the ease of implementation, accuracy, capacity, consumption and size, among others. Based on the case study, we present qualitative directions to choose the most suitable reconfigurable device for similar applications.

A study of topology in insular Parallel Genetic Algorithms

In this paper we study how the connectivity affects the performance of insular Parallel Genetic Algorithms (PGAs). Seven topologies PGAs were proposed, with growing number of connections. We used three instances of the well-known traveling salesman problem as benchmark. Each island of the PGA had different parameters and we established a fixed migration policy for all islands. Experiments were done and average results were reported. The effect of coevolution in PGAs was evidenced. The convergence time increased with the number of connections of the topology. The quality of solutions also increased in the same way. Although topologies with large connectivity increases the overall processing time, they take benefits to the quality of solutions found.

A Compact Genetic Algorithm with Elitism and Mutation Applied to Image Recognition

The problem of object recognition in images is a hard problem frequently found in industrial and academic application. This work presents the application of an extension of the Compact Genetic Algorithm (emCGA) to three problems of object recognition in real images. Results are compared with an exhaustive search algorithm and another CGA. Results suggested the efficiency of emCGA for this problem and encourages future developments.

A Comparative Study of Machine Learning Methods for Detecting Promoters in Bacterial DNA Sequences

Machine Learning methods have been widely used in bioinformatics, mainly for data classification and pattern recognition. The detection of genes in DNA sequences is still an open problem. Identifying the promoter region laying prior the gene itself is an important aid to detect a gene. This paper aims at applying several Machine Learning methods to the construction of classifiers for detection of promoters in the DNA of Escherichia coli. A thorough comparison of methods was done. In general, probabilistic and neural network-based methods were those that performed better regarding accuracy rate.

Implementation of a Parallel Algorithm for Protein Pairwise Alignment Using Reconfigurable Computing

This paper describes the design and implementation of a parallel pairwise alignment hardware, implemented in a FPGA device. This system is aimed at aligning pairs of proteins, using a dynamic programming algorithm. The alignment is done in parallel thanks to a pipelined approach. All functional blocks are described in detail. Experiments were done to access the performance of the system for up to pairs of 2000-amino acids-long proteins. Hardalign was compared with a similar algorithm implemented in software and running in a PC, resulting in a 1:10 speed-up ratio. Results encourage the continuity of the work, showing that reconfigurable computing can offer interesting solutions for bioinformatics problems

An Electric Energy Quality Meter Using Hardware Reconfigurable Computing

The use of electronic loads in production processes has been increased lately in the industries, aiming at improving automation and products quality. At the same time, operational costs are reducing, including the cost of electric energy. Together with these advantages many problems appear, such as electromagnetic interferences and harmonics, justifying the insertion of quality and amount of electric energy meters (EEQ meter) for evaluations and analyses for both, consumers and companies. This work describes and analyses several issues in the implementation of a EEQ meter, emphasizing the functional block implementation using hardware reconfigurable computation. The advantages of this approach, such as execution time and parallelism, are explored in this paper

Construction and improvement heuristics applied to the capacitated vehicle routing problem

The capacitated vehicle routing is a combinatorial optimization problem of that has aroused major interest because it is present in critical areas (logistics, transport and other) and it is a problem of considerable difficulty. There are currently several techniques that have been developed to try to solve this problem efficiently. In this paper, we present a solution to this problem based on the strategy of different single techniques combination to obtain the best results. Computing experiments have been conducted on six instances of well-known data sets available in literature.

Reconfigurable hardware computing for accelerating protein folding simulations using the harmony search algorithm and the 3D-HP-side chain model

Proteins are essentials to life and they have countless biological functions. They are synthesized in the ribosome of cells following a template given by the messenger RNA (mRNA). During the synthesis, the protein folds into an unique three-dimensional structure, known as native conformation. This process is called protein folding. Several diseases are believed to be result of the accumulation of ill-formed proteins.Therefore, understanding the folding process can lead to important medical advancements and development of new drugs.