Fahlberg - The Light (Original Mix)

In principle, our protein engineering framework is applicable whenever an underlying fitness landscape can be inferred via machine learning. There have been multiple previous studies demonstrating the effectiveness of machine learning to navigate the sequence-function landscape. A notable example used a similar UCB method to optimize cytochrome P450 thermostability14. A lower confidence bound (LCB) algorithm was used to predict chimeric channelrhodopsin sequences that localize to the plasma membrane of mammalian cells, and UCB optimization was then used to identify chimeras with high localization18. GP classification and regression models were further used to engineer highly light sensitive channelrhodopsins for optogenetics31. Iterative searches through protein sequence-function landscapes such as UCB optimization and LCB minimization reduce dependency on large datasets, and enable engineering of more difficult protein targets.

Fahlberg - The Light (Original Mix)

The first UCB round was performed slightly differently than the others because we were still refining our method. For the first UCB round, we trained GP regression models on alcohol titers from both BL21(DE3) and CM24 strains. We chose sequences that maximized the sum of the BL21(DE3) and CM24 UCB scores and selected a panel of ten chimeras using the batch mode UCB approach described above. 041b061a72