MDFF_NM: Improved Molecular Dynamics Flexible Fitting into Cryo-EM Density Maps with a Multireplica Normal Mode-Based Search

Molecular Dynamics Flexible Fitting (MDFF) is a widely
used tool
to refine high-resolution structures into cryo-EM density maps. Despite
many successful applications, MDFF is still limited by its high computational
cost, overfitting, accuracy, and performance issues due to entrapment
within wrong local minima. Modern ensemble-based MDFF tools have generated
promising results in the past decade. In line with these studies,
we present MDFF_NM, a stochastic hybrid flexible fitting algorithm
combining Normal Mode Analysis (NMA) and simulation-based flexible
fitting. Initial tests reveal that, besides accelerating the fitting
process, MDFF_NM increases the diversity of fitting routes leading
to the target, uncovering ensembles of conformations in closer agreement
with experimental data. The potential integration of MDFF_NM with
other existing methods and integrative modeling pipelines is also
discussed.