Understanding how transcription regulation is encoded in the genomes of complex multicellular organisms has been a big challenge, not least due to the large non-coding space where relevant interactions might occur. High throughput technologies now allow it to map putative regulatory regions via their chromatin structure, and have made rapid progress in identifying in vivo binding of transcription factors to DNA at high resolution. Large collections of relevant data have been made available by individual groups as well as large consortia such as mod/ENCODE. I will discuss some of our recent and ongoing efforts that make use of such datasets to define successful computational models for individual sites as well as for cell-type specific expression.