Obtaining new genetic regulatory elements has tremendous usage in metabolic engineering and synthetic biology applications. This work introduced an artificial intelligence (AI) framework to create new promoter sequences in E. coli. The authors trained adversarial deep neural networks to learn critical features from natural sequences, and then used the model to generate brand new promoters. These AI designed sequences showed large difference from the E. coli natural genome, but a high proportion of them were demonstrated to be functional. In theory, the model can generate thousands or even millions of new promoters. This work demonstrated the potential ability of AI to explore the huge potential combinations of nucleotide sequences in silico to obtain new optimized genetic elements.
NAR’s Breakthrough Articles present high-impact studies answering long-standing questions in the field of nucleic acids research and/or opening up new areas and mechanistic hypotheses for investigation. These articles are chosen by the Editors on the recommendation of Editorial Board Members and Referees. Articles are accompanied by a brief synopsis explaining the findings of the paper and where they fit in the broader context of nucleic acids research. They represent the very best papers published at NAR.
Bradley et al. have carried out a thorough mechanistic analysis of a previously uncharacterised DNA repair enzyme from E. coli, YcaQ. The protein belongs to a family of DNA glycosylases, which cleave the bond between a damaged base and the DNA backbone. YcaQ is special in that it specifically recognises bases that form part of an interstrand crosslink (ICL), one of the most dangerous types of lesions. This type of activity has been observed before, but is not well understood mechanistically. In their study, they demonstrate that YcaQ defines an ICL repair pathway alternative to such systems as the Fanconi anaemia or NER-based pathways.
Understanding the coupling between steps in gene expression is currently one of the grand challenges in biology. This paper provides strong evidence that the ATP-dependence of pre-spliceosome formation (which is dispensable) is a key player in the coupling between transcription and splicing. Why pre-spliceosome formation is ATP-dependent has been a perplexing question for many years, and this paper provides further support for the notion that this requirement is due to coupling. These authors’ findings also suggest that parts of the core splicing machinery evolved functions not related to intron-removal per se, but rather to facilitate co-transcriptional pre-mRNA processing. Moreover, this coupling may impact both splicing fidelity and the production of intron-less genes in yeast.
