An investigation into the consistency and usability of selected minisatellite detecting software packages

Abstract

A tandem repeat is a sequence of adjacent repetitions of a nucleotide patternsignature, called its motif, in a DNA sequence. The repetitions may either be exact or approximate copies of the motif. A minisatellite is a tandem repeat whose motif is of moderate length. One approach to searching for minisatellites assumes prior knowledge about the motif. This approach limits the search for minisatellites to specified motifs. An alternative approach tries to identify signatures autonomously from within a DNA sequence. Several different algorithms that use this approach have been developed. Since they do not use pre-specified motifs, and since a degree of approximation is tolerated, there may be ambiguity about where minisatellites start and end in a given DNA sequence. Various experiments were conducted on four well-known software packages to investigate this conjecture. The software packages were executed on the same data and their respective output was compared. The study found that the selected computer algorithms did not report the same outputs. The lack of precise definitions of properties of such patterns may explain these differences. The difference in definitions relate to the nature and extent of approximation to be tolerated in the patterns during the search. This problem could potentially be overcome by agreeing on how to specify acceptable approximations when searching for minisatellites. Some of these packages are implemented as Academic/Research Software (ARS). Noting that ARS has a reputation of being difficult to use, this study also investigated the usability of these ARS implementations. It relied on literature that offers usability evaluation methods. Potential problems that are likely to affect the general usability of the systems were identified. These problems relate inter alia, to visibility, consistency and efficiency of use. Furthermore, usability guidelines in the literature were followed to modify the user interface of one of the implementations. A sample of users evaluated the before- and after versions of this user interface. Their feedback suggests that the usability guidelines were indeed effective in enhancing the user interface.