Our library of varied-strength promoters can be utilized in limitless applications of genetic circuitry. Promoters with sequence variability prevent the spontaneous occurrence of homologous recombination within multi-gene cassettes. Given the high usage of the J23119 promoter family, there is significant demand for well-characterized promoters of different strengths for rational engineering of gene expression.
All of the findings generated from the training experiments are immediately useful in their actual projects, rather than being an independent exercise with no external benefit. As such, the exercise was not only an introduction to the engineering, troubleshooting, and analytical parts of their team’s overarching project, but also generated novel code, analysis, and findings that helped shape the trajectory of their project.
The CheRMiT team generated novel, simple implementations of chemical validators in a new library, RDKiT, that informed their existing implementation and revamped it to utilize the functionality of this new Python package; they also expanded the suite of labeling functions useful in generating semi-supervised annotated training data for the training of the project’s customized language models. Both of these resources serve not only as a useful introduction for anyone to become familiar with libraries ranging from RDKit to PyTorch and Snorkel, but have been contributory to important infrastructural and testing components of the project.
The PPI team and HSF team both generated candidate biomolecule sequences, that they then learned how to analyze while becoming familiar with such analytical methods in the process. The candidate sequences were useful in serving as a baseline for future engineering and sequence design encountered in their project, while also equipping students with the familiarity and expertise with various metrics necessary to thoroughly evaluate such candidates through only in silico methods, such as alanine scanning, secondary structural analysis, and binding energy.