Gaining metabolic insights, for novel applications.

Problem description

Artificial intelligence has found its way into microbial sequence analysis. Large amounts of DNA and RNA sequences of microbes and microbial communities are read, with widespread applications in health, food, industrial biotech and agriculture.

Recently, machine learning and AI have been successfully applied in several domains relating to biotechnology, such as structure and function prediction.

References

Host genetics and the rumen microbiome jointly associate with methane emissions in dairy cows

Difford, Gareth Frank, et al.
A heritable subset of the core rumen microbiome dictates dairy cow productivity and emissions

Wallace, R. John, et al.

Goal

To evaluate and implement so-called “Representation learning” methods on multi-omics data from strain development projects - typically in strain-samples series of 100 to 1000 in the near future – to guide experimental design for next generations of strain development. Strain-samples can be from mutagenesis experiments or more targeted rational engineering projects, where genomes, mRNA and/or protein data are available. In first instance, methods will be applied to generate insight in phenotypic behavior based on the identified mutations and/or rationally designed experiments. Secondly, the method will be applied to design novel experiments to build superior combinations of the learned characteristics, either by combinatorial approaches or de-novo design of DNA elements.