SIEVE: Structural Identification and Extraction Via Embedded graphs

Multiscale modeling plays a critical role in understanding complex material systems by linking atomistic phenomena to macroscopic behavior. SIEVE is an open-source C++ tool designed to automate the extraction of meaningful microstates from molecular dynamics (MD) simulations. SIEVE leverages graph theory to reconstruct molecular connectivity directly from Cartesian coordinates and atomic numbers, enabling fast identification, classification, and indexing of molecular species across large-scale MD trajectories. This approach allows for microsecond-per-snapshot analysis of systems with tens of thousands of atoms, maintaining low memory usage and eliminating manual bottlenecks. The core functionality of SIEVE is based on decomposing the simulation box into disjoint molecular graphs, performing graph isomorphism checks against reference structures, and handling efficient I/O through the standard high-performance H5MD format[1][2][3]. Its capabilities straightforwardly apply to the investigation of solvated environments and the formation of molecular aggregates, and may open new possibilities for structural and statistical analysis in complex condensed-phase systems. SIEVE has been successfully applied in three representative cases: first, it has been used to extract microstates from MD trajectories of light-emitting compounds, which have been used for TD-DFT simulations to compute the electronic transitions to gain insight on the possible emission mechanism; second, it has been used to carry out an unsupervised proximity analysis to study molecular aggregation in photoactive solutes solvated in ethanol; and third, it was instrumental in simulating absorption spectra for solvated indigo by selecting solvation environments for TD-DFT calculations. These applications demonstrate the potential of SIEVE as a transformative tool for multiscale modeling workflows. [1] de Buyl, P., Colberg, P. H., & Höfling, F. (2014). H5MD: A structured, efficient, and portable file format for molecular data. Computer Physics Communications, 185(6), 1546-1553, https://doi.org/10.1016/j.cpc.2014.01.018. [2] https://www.boost.org/doc/libs/1_83_0/libs/graph/doc/index.html [3] McKay, B.D. and Piperno, A., Practical Graph Isomorphism, II, Journal of Symbolic Computation, 60 (2014), pp. 94-112, https://doi.org/10.1016/j.jsc.2013.09.003.