Genome-wide detection of alternative splicing in expressed sequences of human genes

We have identified 6201 alternative splice relationships in human genes, th rough a genome-wide analysis of expressed sequence tags (ESTs), Starting wi th similar to2.1 million human mRNA and EST sequences, we mapped expressed sequences onto the draft human genome sequence and only accepted splices th at obeyed the standard splice site consensus. A large fraction (47%) of the se were observed multiple times, indicating that they comprise a substantia l fraction of the mRNA species. The vast majority of the detected alternati ve forms appear to be novel, and produce highly specific, biologically mean ingful control of function in both known and novel human genes, e.g. specif ic removal of the lysosomal targeting signal from HLA-DM beta chain, replac ement of the C-terminal transmembrane domain and cytoplasmic tail in an FC receptor beta chain homolog with a different transmembrane domain and cytop lasmic tail, likely modulating its signal transduction activity. Our data i ndicate that a large proportion of human genes, probably 42% or more, are a lternatively spliced, but that this appears to be observed mainly in certai n types of molecules (e.g. cell surface receptors) and systemic functions, particularly the immune system and nervous system. These results provide a comprehensive dataset for understanding the role of alternative splicing in the human genome, accessible at http://www.bioinformatics.ucla.edu/HASDB.