Simulating 500 million years of evolution with a language model.

More than three billion years of evolution have produced an image of biology encoded into the space of natural proteins. Here we show that language models trained at scale on evolutionary data can generate functional proteins that are far away from known proteins. We present ESM3, a frontier multimodal generative language model that reasons over the sequence, structure, and function of proteins.

Limited Migration From Physiological Refugia Constrains the Rescue of Native Gastropods Facing an Invasive Predator.

Biological invasions have caused the loss of freshwater biodiversity worldwide. The interplay between adaptive responses and demographic characteristics of populations impacted by invasions is expected to be important for their resilience, but the interaction between these factors is poorly understood. The freshwater gastropod is native to the Upper St.

Prospector Heads: Generalized Feature Attribution for Large Models & Data.

Feature attribution, the ability to localize regions of the input data that are relevant for classification, is an important capability for ML models in scientific and biomedical domains. Current methods for feature attribution, which rely on "explaining" the predictions of end-to-end classifiers, suffer from imprecise feature localization and are inadequate for use with small sample sizes and high-dimensional datasets due to computational challenges. We introduce prospector heads, an efficient and interpretable alternative to explanation-based attribution methods that can be applied to any encoder and any data modality.

Designing eco-evolutionary experiments for restoration projects: Opportunities and constraints revealed during stickleback introductions.

Eco-evolutionary experiments are typically conducted in semi-unnatural controlled settings, such as mesocosms; yet inferences about how evolution and ecology interact in the real world would surely benefit from experiments in natural uncontrolled settings. Opportunities for such experiments are rare but do arise in the context of restoration ecology-where different "types" of a given species can be introduced into different "replicate" locations. Designing such experiments requires wrestling with consequential questions.