“…Genes from each CEM were first mapped to the identified operons to retrieve the basic transcription units. Next, 300 bps in the upstream of the translation starting sites for each operon were extracted, in which motif finding was carried out using the web server DMINDA (Ma, Zhang, et al 2014, Yang et al 2017). DMINDA is a dominant motif prediction tool, embraced five analytical algorithms to find, scan, and compare motifs (Li, Liu, Ma and Xu 2011, Liu et al 2017, Ma, Liu, Zhou, Yin, Li and Xu 2013), including a phylogenetic footprint framework to elucidate the mechanism of transcriptional regulation at a system level in prokaryotic genomes (Li, Ma, Mao, Yin, Zhu and Xu 2011, Liu, Zhang, Zhou, Li, Fennell, Wang, Kang, Liu and Ma 2016, Liu, Zhou, Li, Zhang, Zeng, Liu and Ma 2016).…”