DeepMind presented remarkably accurate predictions at the recent CASP14 protein structure prediction assessment conference. We explored network architectures incorporating related ideas and obtained the best performance with a three-track network in which information at the 1D sequence level, the 2D distance map level, and the 3D coordinate level is successively transformed and integrated. The three-track network produces structure predictions with accuracies approaching those of DeepMind in CASP14, enables the rapid solution of challenging X-ray crystallography and cryo-EM structure modeling problems, and provides insights into the functions of proteins of currently unknown structure. The network also enables rapid generation of accurate protein-protein complex models from sequence information alone, short circuiting traditional approaches which require modeling of individual subunits followed by docking. We make the method available to the scientific community to speed biological research.

Saturation mutagenesis probes define sections of the vast protein sequence space. However, even if randomization is limited this way, the combinatorial numbers problem is severe. Because diversity is created at the codon level, codon redundancy is a crucial factor determining the necessary effort for library screening. Additionally, due to the probabilistic nature of the sampling process, oversampling is required to ensure library completeness as well as a high probability to encounter all unique variants. Our trick employs a special mixture of three primers, creating a degeneracy of 22 unique codons coding for the 20 canonical amino acids. Therefore, codon redundancy and subsequent screening effort is significantly reduced, and a balanced distribution of codon per amino acid is achieved, as demonstrated exemplarily for a library of cyclohexanone monooxygenase. We show that this strategy is suitable for any saturation mutagenesis methodology to generate less-redundant libraries.

A series of synthetic nicotinamide cofactors were synthesized to replace natural nicotinamide cofactors and promote enoate reductase (ER)-catalyzed reactions without compromising activity or stereoselectivity of the bioreduction process. Conversions and enantioselectivities of up to >99% were obtained for C=C bioreductions and the process was successfully upscaled. Furthermore, high chemoselectivity was observed when employing these nicotinamide cofactor mimics (mNADs) with crude extracts in ER-catalyzed reactions.The asymmetric reduction of conjugated C=C double bonds using enoate reductases (ERs, EC 1.3.1.31) is receiving great interest in preparative organic chemistry.1 § Both authors contributed equally.