PT  - JOURNAL ARTICLE
AU  - Shiori Sagawa
AU  - Morgan N. Price
AU  - Adam M. Deutschbauer
AU  - Adam P. Arkin
TI  - Validating Regulatory Predictions from Diverse Bacteria with Mutant Fitness Data
AID  - 10.1101/091405
DP  - 2016 Jan 01
TA  - bioRxiv
PG  - 091405
4099  - http://biorxiv.org/content/early/2016/12/06/091405.short
4100  - http://biorxiv.org/content/early/2016/12/06/091405.full
AB  - Although transcriptional regulation is fundamental to understanding bacterial physiology, the targets of most bacterial transcription factors are not known. Comparative genomics has been used to identify likely targets of some of these transcription factors, but these predictions typically lack experimental support. Here, we used mutant fitness data, which measures the importance of each gene for a bacterium’s growth across many conditions, to validate regulatory predictions from RegPrecise, a curated collection of comparative genomics predictions. Because characterized transcription factors often have correlated fitness with one of their targets (either positively or negatively), correlated fitness patterns provide support for the comparative genomics predictions. At a false discovery rate of 3%, we identified significant cofitness for at least one target of 158 TFs in 107 ortholog groups and from 24 bacteria. Thus, high-throughput genetics can be used to identify a high-confidence subset of the sequence-based regulatory predictions.