Base Qualities Help Sequencing Software

With the complete sequencing of the human genome under way and the sequencing of complete microorganism genomes becoming commonplace, we have truly entered the era of large-scale DNA sequencing. Why now? As in some other data-rich areas of modern biology, for example, protein structure determination, it can be argued that the rate-limiting factors in increasing efficiency and throughput have been computer power and software. We could have run thousands of sequencing gels 20 years ago, but without image-processing software and fragment assembly packages it would not have been feasible to put together all of the individual sequence fragments from the gels to give megabases of continuous, accurate sequence. At any rate, the development of powerful computational tools is central to large-scale sequencing.

This special informatics issue contains several papers on the software used in genome sequencing centers, and in particular three papers on the set of programs from Phil Green’s group at the University of Washington in Seattle (Ewing and Green 1998; Ewing et al. 1998; Gordon et al. 1998). These programs have played a key role in the progress of the largest-scale projects under way. They have been used extensively in the 100-Mb Caenorhabditis elegans project being completed this year and predominate among groups sequencing the human genome.

Such sequencing groups start with large clones such as BACs or PACs of 100 kb or more, or small genomes of up to a few megabases, for which the goal is to obtain complete accurate sequence. However, the raw sequences, or “reads,” obtained from the gels run on automated machines such as ABI 377s are only on the order of 500–1000 bp …