Search bioRxiv⌕ Search

Biology subjects

Wahlberg, M.

Publications and source records attributed to Wahlberg, M..

2 recordsLinked to original sources

Otters, but not apes, prepare for mutually exclusive possibilities.

The ability to prepare for mutually exclusive outcomes is often considered uniquely human. Solving such problems requires anticipating alternative futures before acting. In the classic forked-tube task, the optimal strategy is to block both exits to secure a reward: children under four years and great apes typically fail, whereas older children succeed. Using this paradigm, we tested three Asian small-clawed otters (Aonyx cinereus) and one Eurasian otter (Lutra lutra) and compared their performance with chimpanzee data. Otters covered both exits significantly more often than chimpanzees, with all individuals succeeding within their first trials. Initial inconsistency in maintaining the strategy appeared linked to anatomical constraints that limited reward success. When retested two months later with an apparatus better suited to otter morphology, individuals adopted and maintained dual coverage as success increased, indicating that the behavior tracked the payoff structure of the task rather than reflecting low-level mechanisms such as trial-and-error learning. Together, these findings indicate that blocking both exits is an adaptive response to the tasks causal structure, supporting the ecological intelligence hypothesis: cognition evolves in response to ecological demands, particularly foraging challenges that place recurrent pressures on memory, decision-making, and executive control, rather than being driven solely by social complexity.

animal behavior and cognition↗

Integration of variant annotations using deep set networks boosts rare variant association genetics

Rare genetic variants can strongly predispose to disease, yet accounting for rare variants in genetic analyses is statistically challenging. While rich variant annotations hold the promise to enable well-powered rare variant association tests, methods integrating variant annotations in a data-driven manner are lacking. Here, we propose DeepRVAT, a model based on set neural networks that learns burden scores from rare variants, annotations, and phenotypes. In contrast to existing methods, DeepRVAT yields a single, trait-agnostic, nonlinear gene impairment score, enabling both risk prediction and gene discovery in a unified framework. On 34 quantitative and 26 binary traits, using whole-exome-sequencing data from UK Biobank, we find that DeepRVAT offers substantial increases in gene discoveries and improved replication rates in held-out data. Moreover, we demonstrate that the integrative DeepRVAT gene impairment score greatly improves detection of individuals at high genetic risk. Finally, we show that pre-trained DeepRVAT scores generalize across traits, opening up the possibility to conduct highly computationally efficient rare variant tests.

bioinformatics↗