See our poster for the quarter summary here
UW-DIRECT Project on metabolite retrosynthetic analysis
- Aim to utilize data science and software engineering intuition to find, and predict, a plausible metabolic pathway for production of a given molecule with retrosynthetic analysis approach.
- Current focus is to find a novel promiscuous substrate for enzymatic transformation.
- Add compounds with canonical SMILES string but no isomeric SMILES string
- Test inclusion of additional features, such as full chemical fingerprints and enzyme descriptors
- Explore alternatives models such as SVMs, neural networks, decision trees/random forests, and ensemble methods
- Extend approach to include non-promiscuous enzymes
- Include simple chemical transformation for biocatalysis application
For thorough explanation, use cases, and project background --- Check out our Wiki
- BioPython
- PubChemPy
- RDKit
- pandas
- sklearn
- scipy
- numpy
