The application potential of machine learning and genomics for understanding natural product diversity, chemistry, and therapeutic translatability

Abstract

The natural product field continues to benefit from machine learning, both chemically and biologically. Together machine learning, genomics, and natural products have great potential for improving drug discovery and impacting human health.

Alkaloids in Contemporary Drug Discovery to Meet Global Disease Needs

“…At the present, a complete natural product structure cannot be predicted de novo from a new biosynthetic gene cluster; although algorithms may assist [274], heterologous expression of the BGC and conducting the isolation and structure elucidation is required. Once non-functional modules can be unambiguously identified in silico, and substrate specificities defined, levels of structure prediction for an assembled pathway product will be enhanced.…”

A roadmap for metagenomic enzyme discovery

“…There are a growing number of BGC-specific machine learning tools available to predict natural product structure and bioactivity from metagenomes. These include BGC detection and classification software reviewed elsewhere 187 such as antiSMASH, 94 PRISM, 188 DeepBGC 189 and most recently, GECCO ( ). Unfortunately, researchers tend to work either on the more general protein function prediction problem or on natural products biosynthesis, but they do not often communicate with each other.…”

Natural Products, the Fourth Industrial Revolution, and the Quintuple Helix

“…92 Identifying the gene cluster does not predicate a natural product structure. Although several algorithms are available to assist, 93 heterologous expression, isolation, and structure elucidation are still required at the present time. As more gene clusters are characterized and the substrate specificities recognized, higher levels of predictive accuracy should be expected, particularly if non-functional modules within a gene product in an assembly sequence can be identified.…”

“…247 The ability to adjust and "learn" based on additional, perhaps modulated, data input in a feedback loop, and seek new patterns for recognition provides creative insights into the chemical and biological relationships of natural products. 93 Within the natural product sciences, the potential applications are extensive and include recognition of genomic signature elements and predictions about collective outcomes biosynthetically, projections of bioactivity, propositions for (bio)synthetic compound diversity, and disease targeting. 248 Deep learning (DL) is an extension of ML and focuses on layers of neural networks that can assist in predicting protein structures.…”