Aron Stålmarck

This paper shows a transformer-based generative model for proteins spontaneously creates symmetry without explicit training or generation constraints.

This paper introduces CHMMAIRRa, a hidden Markov model to detect PCR chimeras in AIRR-seq data, incorporating somatic hypermutation and germline references.

This paper introduces a generative autoregressive language model for protein backbone design, showing it can sample diverse and realistic protein chains.