Quantitative modeling of a genes expression from its intergenic sequence. Md Abul Hassan Samee¹, Tara Lydiard-Martin², Angela DePace², Saurabh Sinha^1,3. 1) Dept of Comp Sci, Univ of Illinois, Urbana, IL; 2) Dept of Systems Biology, Harvard Medical School, Boston, MA; 3) Institute for Genomic Biology, Univ of Illinois, Urbana, IL.

   Modeling a genes expression from its intergenic locus is a fundamental goal in computational biology. Owing to the distributed nature of cis-regulatory information and the poorly understood mechanisms that integrate such information, gene locus modeling is a more challenging task than modeling an individual enhancer. Here we report the first quantitative model of a genes expression pattern as a function of its intergenic locus. We model the expression readout of a locus in two tiers: 1) combinatorial effect of transcription factor (TF) binding sites within each enhancer is predicted by a thermodynamics-based model and 2) contributions from enhancers are linearly combined to fit the gene expression pattern. The model works without any prior knowledge of the locations of enhancers. We show that the model accurately fits the expression patterns of the genes eve, h, run, and gt along the A/P axis of the Drosophila embryo. Our results suggest that there are sequence segments, located in inaccessible regions of the locus, that have ectopic expression readouts and were explicitly avoided by the model. We applied our model to identify the TFs forming the stripe boundaries of the studied genes. The resulting networks of TF-stripe relations show remarkable agreement with the known regulatory influences on these genes. We also developed a computational framework using the model to probe for interactions between enhancers of a gene. We found that enhancers of our studied genes tend to act autonomously to drive the respective gene expression patterns. Finally, we show that our model is able to explain the readouts from a novel set of constructs where the eve 3/7 and the eve 4/6 enhancers were fused in different orientations and with different spacers at the junctions, while existing models of enhancer function failed to do so.