CheSweet: An application to predict glycans chemicals shifts

GARAY PABLO G.; ARROYUELO AGUSTINA; VILA JORGE A.; MARTIN OSVALDO O.

San Luis Capital

Congreso; PyData San Luis; 2017

Instituto de Matemática Aplicada San Luis

Glycans are the most abundant and structurally diverse biomolecules in nature. The knowledge of the tridimensional structure of these molecules is necessary to understand in detail, at atomic level, the molecular processes in which glycans are involved. Chemical shifts (CS) are observables obtained from Nuclear Magnetic Resonance experiments and are highly sensitive probes to sense conformational changes. Here we present CheSweet a Python module to compute CS for glycans. The core of CheSweet is the fast calculation of CS based on the pre-calculated values of CS, computed at DFT level of theory, from the values of the torsional angles (phi, psi, omega and chi). CS are calculated through linear interpolation or by nearest neighbors when linear interpolation is not applicable. We will discuss the details of the code implementation, and also the use of CheSweet as part of a Bayesian model (using PyMC3) to predict glycan structures.