A BRANCH POINT CONSENSUS FROM ARABIDOPSIS FOUND BY NONCIRCULAR ANALYSIS ALLOWS FOR BETTER PREDICTION OF ACCEPTOR SITES

Little knowledge exists about branch points in plants; it has even bee n claimed that plant introns lack conserved branch point sequences sim ilar to those found in vertebrate introns. A putative branch point con sensus sequence for Arabidopsis thaliana resembling the well known met azoan consensus sequence has been proposed, but this is based on searc h of sequences similar to those in yeast and metazoa. Here we present a novel consensus sequence found by a non-circular approach. A hidden Markov model with a fixed A nucleotide was trained on sequences upstre am of the acceptor site. The consensus found by the Markov model share s features with the metazoan consensus, but differs in its details fro m the consensus proposed earlier. Despite the fact that branch point c onsensus sequences in plants are weak, we show that a prediction schem e incorporating them leads to a substantial improvement in the recogni tion of true acceptor sites; the false positive rate being reduced by a factor of 2. We take this as an indication that the consensus found here is the genuine one and that the branch point does play a role in the proper recognition of the acceptor site in plants.