Copenhagen Bioinformatics Hackathon 2021
Our Team
Icons CC BY SA OpenMoji
Hrishikesh Michael Simon Emzo
@hbdhondge @simonduerr @scyphs
Model architecture
Loss: BCE + KLD + HuberLoss
reconstruction Smooth L1 norm to stability
Model architecture
This is a generative model!
We can sample sequences around a predicted stable or unstable sequence from the latent space.
Single mutant test set
Multi mutant test set
ProtTransBertBFD
Elnaggar, Ahmed, et al. “ProtTrans: Towards Cracking the Language of Life’s Code Through Self-Supervised Deep Learning and High Performance Computing.” arXiv preprint arXiv:2007.06225 (2020). https://arxiv.org/abs/2007.06225
2 Conv layers
1x Dropout (0.25)
Using the same architecture as for the single mutants