Special Issue: High Performance Computational Biology

Ever since the structure of DNA was discovered in 1953, biology has been steadily changing from being a descriptive science concerned with the behavior and characteristics of organisms to a mathematical discipline that relates the essential life processes to the underlying biomolecular data. This discovery has stimulated the growth of molecular biology, the study of how biomolecular sequences are related to the functioning of organisms. These developments have brought biology closer to computer science. In many ways, the underlying mechanisms are similar to what we employ in building and programming computers. The characteristics of a life form are coded in its DNA (program), which is processed in each cell (executed) to produce the proteins (outputs) that carry out essential life processes. The field holds immense potential for future discoveries that are unrivaled in significance such as the design of protein sequences to fold into a specific configuration to efficiently administer drugs and the possibility of treating diseases by altering the genetic code. The need to discover biomolecular sequences, to relate the sequences to their structure and function and to understand the sequences through mutual comparison, has resulted in a number of interesting problems for algorithm designers and led to the development of computational molecular biology. The area has attracted many competent researchers and the field is developing at a rapid pace as evidenced by the growth of conference meetings and avenues for publication. Algorithms for solving biological problems are often associated with long running times. This arises due to various factors. (1) Biological data are obtained by experiments which are prone to errors. The need to deal with errors and uncertainties results in algorithms with high complexity. (2) The data size itself may be large and result in long running times. (3) Many of the problems are shown to be NP-hard and techniques such as energy minimization and branch and bound are used. As biologists progressed from the study of simple biomolecular data from less complex organisms to the eventual goal of understanding and manipulating entire genomes of complex organisms, the corresponding computational needs are scaling similarly. We believe that effective use of parallel computers is becoming increasingly important for solving meaningful biological problems in reasonable time. The field of computational molecular biology is replete with applications that require processing large amounts of data. The basic problem of finding DNA sequences that exhibit homology to a given query sequence requires searching databases containing over tens of billions of nucleotides, and still growing at an exponential rate. The recent assembly of the mouse genome required processing over 33 million fragments of a total size of over 17 billion bases to assemble the genome of size over 3 billion bases. In comparative genomics, two or more genomes of such enormous sizes must