Högskolan i Skövde

As the interest in biological sequence analysis increases, more efficient techniques to sequence, map and analyse genome data are needed. One frequently used technique is EST sequencing, which has proven to be a fast and cheap method to extract genome data. An EST sequencing generates large numbers of low-quality sequences which have to be managed and analysed further.

Performing complete searches and finding guaranteed results are very time consuming. This dissertation project presents a method that can be used to perform rapid gene prediction of function-specific genes in EST data, as well as the results and an estimation of the accuracy of the method.

This dissertation project applies various methods and techniques on actual data, attempting to identify genes involved in cold-associative processes in plants. The presented method consists of three steps. First, a database with genes known to have cold-associated properties is assembled. These genes are extracted from other, already sequenced and analysed organisms. Secondly, this database is used to identify homologues in an unanalysed EST dataset, generating a candidate-list of cold-associated genes. Last, each of the identified candidate cold-associative genes are verified, both to estimate the accuracy of the rapid gene prediction and also to support the removal of candidates which are not cold-associative.

The method was applied to a previously unanalysed Avena sativa EST dataset, and was able to identify 135 candidate genes from approximately 9500 EST's. Out of these, 103 were verified as cold-associated genes.