Background: As research into alternative splicing reveals the fundamental importance of this phenomenon in the genome expression of higher organisms, there is an increasing need for a standardized, consistent and unique identifier for alternatively spliced isoforms. Such an identifier would be useful to eliminate ambiguities in references to gene isoforms, and would allow for the reliable comparison of isoforms from different sources (e.g., known genes vs. computational predictions). Commonly used identifiers for gene transcripts prove to be unsuitable for this purpose.
Methodology: We propose an algorithm to compute an isoform signature based on the arrangement of exons and introns in a primary transcript. The isoform signature uniquely identifies a transcript structure, and can therefore be used as a key in databases of alternatively spliced isoforms, or to compare alternative splicing predictions produced by different methods. In this paper we present the algorithm to generate isoform signatures, we provide some examples of its application, and we describe a web-based resource to generate isoform signatures and use them in database searches.
Conclusions: Isoform signatures are simple, so that they can be easily generated and included in publications and databases, but flexible enough to unambiguously represent all possible isoform structures, including information about coding sequence position and variable transcription start and end sites. We believe that the adoption of isoform signatures can help establish a consistent, unambiguous nomenclature for alternative splicing isoforms. The system described in this paper is freely available at http://genome.ufl.edu/genesig/, and supplementary materials can be found at http://genome.ufl.edu/genesig-files/.