# Low-N Protein Activity Optimization with FolDE

arXiv: https://arxiv.org/abs/2510.24053 · HTML: https://mlsanity.com/paper/2510.24053/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 8 of its 35 references that are in the map.
- First posted on arXiv 28 Oct 2025.

## Builds on (most cited first, 8 of 8)

- [A Survey of Uncertainty in Deep Neural Networks](https://mlsanity.com/paper/2107.03342/) (2021)
- [The Parallel Knowledge Gradient Method for Batch Bayesian Optimization](https://mlsanity.com/paper/1606.04414/) (2016)
- [Virology Capabilities Test (VCT): A Multimodal Virology Q&A Benchmark](https://mlsanity.com/paper/2504.16137/) (2025)
- [Adaptive machine learning for protein engineering](https://mlsanity.com/paper/2106.05466/) (2021)
- [Kermut: Composite kernel regression for protein variant effects](https://mlsanity.com/paper/2407.00002/) (2024)
- [ODBO: Bayesian Optimization with Search Space Prescreening for Directed Protein Evolution](https://mlsanity.com/paper/2205.09548/) (2022)
- [Fine-tuning Protein Language Models with Deep Mutational Scanning improves Variant Effect Prediction](https://mlsanity.com/paper/2405.06729/) (2024)
- [Contrastive losses as generalized models of global epistasis](https://mlsanity.com/paper/2305.03136/) (2023)

## Abstract (arXiv)

Proteins are traditionally optimized through the costly construction and measurement of many mutants. Active Learning-assisted Directed Evolution (ALDE) alleviates that cost by predicting the best improvements and iteratively testing mutants to inform predictions. However, existing ALDE methods face a critical limitation: selecting the highest-predicted mutants in each round yields homogeneous training data insufficient for accurate prediction models in subsequent rounds. Here we present FolDE, an ALDE method designed to maximize end-of-campaign success. In simulations across 20 protein targets, FolDE discovers 23% more top 10% mutants than the best baseline ALDE method (p=0.005) and is 55% more likely to find top 1% mutants. FolDE achieves this primarily through naturalness-based warm-starting, which augments limited activity measurements with protein language model outputs to improve activity prediction. We also introduce a constant-liar batch selector, which improves batch diversity; this is important in multi-mutation campaigns but had limited effect in our benchmarks. The complete workflow is freely available as open-source software, making efficient protein optimization accessible to any laboratory.