Publications by authors named "Samantha R Sgrizzi"

Directed protein evolution is central to biomedical applications but faces challenges like experimental complexity, inefficient multi-property optimization, and local maxima traps. While methods using protein language models (PLMs) can provide modeled fitness landscape guidance, they struggle to generalize across diverse protein families and map to protein activity. We present EVOLVEpro, a few-shot active learning framework that combines PLMs and regression models to rapidly improve protein activity.

View Article and Find Full Text PDF
Article Synopsis
  • Directed evolution of proteins is crucial for various fields but is traditionally labor-intensive and limited in efficiency.
  • The introduction of EVOLVEpro, a few-shot active learning framework, enhances protein activity optimization using protein language models and activity predictors, achieving significant improvements in fewer rounds.
  • EVOLVEpro demonstrated substantial advancements across different proteins and applications, potentially transforming AI-guided protein engineering in biology and medicine.
View Article and Find Full Text PDF

Smooth muscle cell (SMC) contraction and vascular tone are modulated by phosphorylation and multiple modifications of the thick filament, and thin filament regulation of SMC contraction has been reported to involve extracellular regulated kinase (ERK). Previous studies in ferrets suggest that the actin-binding protein, calponin 1 (CNN1), acts as a scaffold linking protein kinase C (PKC), Raf, MEK and ERK, promoting PKC-dependent ERK activation. To gain further insight into this function of CNN1 in ERK activation and the regulation of SMC contractility in mice, we generated a novel Calponin 1 knockout mouse (Cnn1 KO) by a single base substitution in an intronic CArG box that preferentially abolishes expression of CNN1 in vascular SMCs.

View Article and Find Full Text PDF

Programmable approaches to sense and respond to the presence of specific RNAs in biological systems have broad applications in research, diagnostics, and therapeutics. Here we engineer a programmable RNA-sensing technology, reprogrammable ADAR sensors (RADARS), which harnesses RNA editing by adenosine deaminases acting on RNA (ADAR) to gate translation of a cargo protein by the presence of endogenous RNA transcripts. Introduction of a stop codon in a guide upstream of the cargo makes translation contingent on binding of an endogenous transcript to the guide, leading to ADAR editing of the stop codon and allowing translational readthrough.

View Article and Find Full Text PDF