PALMA

PALMA is an alignment approach that uses large margin learning to combine splice site predictions with sequence alignment techniques for accurately aligning mRNA sequences to genomic DNA in the presence of sequencing errors, alternative splicing, and micro-exons. PALMA is shown to accurately identify intron boundaries and optimal local alignment boundaries in a carefully designed experiment and outperforms all other methods tested, demonstrating robustness to mutations, insertions, and deletions.