Understanding gene regulation is a central problem in molecular biology. The adoption of large-scale biological data generation techniques such as expression and tiling microarrays has enabled researchers to tackle the gene regulation problem in a global way. Using the yeast as a model system, we explore the combined use of gene upstream sequence signals to explain the observed mRNA variations and to analyze "epi-genetic" effects of histone acetylations.
We repeated the analysis of Beer and Tavazoie (2004, Cell), and observed that sequence motifs indeed have a sigificiant predictive power of
gene expression changes. However, their "Bayesian netowrk" models tend
to overfit and are not globally supported by the data, and their cross-validation studies were biased in favor of their claims. By further using the 666 sequence motifs obtained by Beer and Tavazoie, together with some recent histone acetylation data, we found that the
regulatory roles of H3 and H4 acetylations are different and that their epigenetic effects on gene expression is significant even after one
accounts for the sequence motif signals.
Copyright. All Rights Reserved.