RECOGNITION OF GENES IN DNA-SEQUENCE WITH AMBIGUITIES

The search for genes in a newly sequenced DNA is a well known problem. Among other factors, the gene-searching process is hampered by a numb er of ambiguities which may remain unresolved experimentally for a lon g time. A computer method that is able to predict genes in a DNA seque nce containing ambiguities has been developed, based on the non-homoge neous Markov chain technique. The reliability of the method has been t ested using a set of sequences generated by a Monte-Carlo procedure an d a set of 425 E. coli sequences with ambiguities introduced artificia lly.