Almost every messenger RNA contains a natural long "tail" -- a strand of hundreds of adenosine bases located on one end. The conventional method being used in past decades converts each messenger RNA into DNA molecule, which also includes a long "tail" copied from the messenger RNA.
During the subtraction reaction, tangled hybrids form randomly between these long tails of unrelated DNA molecules. Because all the hybrids are removed after subtraction, many unknown genes in the sample pool become the victims; they are also removed inadvertently even before having a chance to be discovered. The genes expressed at low level are particularly effected by this phenomenon.
The authors of the PNAS paper devised a specific method to truncate these troublesome tails, reducing them from hundreds down to tens of adenosine bases. They showed that removal of the tails resulted in the retention of 1.4 to 7.8 times higher number of copies in four out of five colon-specific genes expressed at low levels after subtraction.
When they applied their method to fish out genes in a previously well-characterized sample from colon cells, they found that many unknown genes were still hidden within that sample and these genes can be identified with their new method.
"This could make a real difference in the increasingly difficult process of cataloguing the human genes," said Janet Rowley, M.D., Blume-Reise Distinguished Service Professor in the departments of medicine, molecular genetics & cell biology, and human genetics at the University of Chicago and director of the study. "It could also simplify the postgenome process of learning what these less common genes do and when they are expressed."
Additional authors of the study include Scott Fears and Jian-Jun Chen of the University of Chicago and Lin Zhang of Johns Hopkins.