Probing coexistence of robust threshold and ultrasensitivity in molecular switches and cascades

Rushmila Shehreen Khan, Md. Shahriar Karim

Keywords: Molecular switches, Stochastic noise, PINN-based model

TL;DR: Robust threshold and ultrasensitivity in molecular switches, cascades and PINN-based predictability

Abstract: Molecular switches are fundamental building blocks in many biophysical processes. These switches arise from dissimilar molecular interactions, affected by intrinsic and extrinsic noise, but exhibit a common ON-OFF behavior that suffices for signal transduction, amplification, and relay. Such repertoires of different mechanisms pose a fundamental question: which mechanisms in a noisy environment confer ultrasensitivity as a switch while also ensuring robust decoding of the concentration threshold required for information interpretation? In this paper, we design molecular switches based on molecular-exchange mechanisms (MEM) and dimeric ligand formation, comprising reversible interactions, and compare them stochastically with other switch-forming mechanisms. Our results reveal that the MEM and Dimer mechanisms exhibit improved robustness of the threshold against noise and appear less uncertain if assessed information-theoretically. This superiority is reflected in their switch-like behavior, making them less prone to flipping states in noisy environments. Extending beyond single-switch comparisons, we also explore alternative switch cascades and assess their performance against noise. Alongside this, the continuing expansion of this work is transformed to a Physics-Informed Neural Network (PINN) form that can predict switching behavior for variation in input signal. More importantly, the PINN-switch can ensure that any agentic-AI driven design adheres to the biological and physical constraints indispensable in a molecular switch. Insights from this study may be helpful in drug design and synthetic circuits aimed at therapeutics and other aqueous applications.

Email Sharing: We authorize the sharing of all author emails with Program Chairs.

Data Release: We authorize the release of our submission and author names to the public in the event of acceptance.

Submission Number: 214