Life sciences have emerged as a primary application area for the use of GPU computing. This is mainly caused by the large amount of publicly available sequence, expression, and structure data. The amount of available data will grow even further in the near future owing to advances in high-throughput technologies leading to a data explosion. Because GPU performance grows faster than CPU performance, the use of GPUs in the life sciences is therefore a perfect match.

A particular area of interest in this context is next-generation sequencing (NGS) technology, which can now produce billions of sequences (reads) on a daily basis. The usage of GPUs can thus play a key role in NGS, and its future applications (such as personal genomics) by providing the necessary computing power to process and analyze this data.

Chapter 11 by Ligowski, Rudnicki, Liu, and Schmidt describes how the popular Smith-Waterman algorithm for protein sequence database scanning can be optimized on GPUs. Starting from a basic CUDA implementation, several optimization techniques using shared memory, registers, loop unrolling, and CPU/GPU partitioning are presented. The combination of these techniques leads to a fivefold performance improvement on the same hardware.

How CUDA can be used to accelerate the folding of an RNA sequence is shown by Rizk, Lavenier, and Rajopadhye in Chapter 14 . The authors achieve a highly efficient CUDA implementation by introducing a reordering of the given sequential algorithm that allows tiled computations and data reuse on the GPU.

In Chapter 12 , Weiss and Bailey explore the process of optimizing mapping of short read data produced by NGS technologies to a reference genome. By using a number of techniques, such as new data layouts, an improvement of two orders of magnitude over the initial CUDA implementation is achieved.

Khajeh-Saeed and Perot describe the computation of a single very large pattern matching search on a large GPU cluster in Chapter 13 . The authors reformulate the Smith-Waterman algorithm in a way that allows the use of parallel scan operations across multiple GPUs, resulting in an efficient implementation even for slow GPU interconnection links.

Finally, in Chapter 15 Feng, Cao, Patnaik, and Ramakrishnan present a solution for mining spike train datasets produced by multielectrode arrays on GPUs. Two strategies for efficiently mapping the problem onto CUDA are described: one thread per occurrence and two-pass elimination.