The forgotten variables of DNA array hybridization

Abstract The reproducibility and reliability of DNA array measurements have been repeatedly questioned during the years. A reassessment of fundamental variables of nucleic acid hybridization might help to solve some of these problems. The hybridization equilibrium, t(1/2), is 41 days for a target human genome. Similar hybridization t(1/2) are expected for whole-transcriptome chips hybridized with tissue cDNA. This implies that most studies on mammalian cells have not been performed under equilibrium conditions. Non-equilibrium binding introduces a stochastic factor into hybridization dynamics; in other words, hybridization will prevail at different spots in different experiments, everything else being equal. A careful re-evaluation of results obtained under non-equilibrium conditions might prove fruitful and help to explain instances where findings conflict.