Systems
We are interested in light-matter interactions in biological systems. The major systems of interest are :
Melanin - Structure and Property
Eumelanin, the functional polymer in human skin, forms a heterogeneous layered structure intrinsic to its broadband monotonic spectra. The inherent structural heterogeneity of eumelanin makes the photoprocesses very complex and diverse in nature. Due to this diversity, a complete mechanistic picture of these photoprocesses, essential to understanding the photoprotective properties, has been missing to date.
The heterogeneity and insolubility in water also makes the structural determination very challenging.
Green Fluorescent Protein Chromophore
​
Interactions with the environment tune the spectral properties of biological chromophores, e.g., fluorescent proteins. Understanding the relative contribution of the various types of noncovalent interactions in the spectral shifts can provide rational design principles toward developing new fluorescent probes.
Furthermore, the photophysics and photochemistry of the chromophores after excitation is fascinating.
Nucleic Acid Bases​
​
Ionization of nucleobases is affected by their biological environment, which includes both the effect of adjacent nucleotides as well as the presence of water around it. We study the effect of environment on the excited state properties of DNA bases.
Methods
Hybrid QM/Effective Fragment Potential
​
The prediction of accurate solvatochromic shifts to the electronic excited states of chromophores is a challenge, especially in the complex biological phase, due to the importance of long-range electrostatic interactions. Hybrid QM/MM methods are generally employed for the calculation of quantum mechanical properties in complex systems. To be predictive, there is a need for an accurate quantum mechanical method that can depict the charge transfer states correctly and incorporate higher than single excited determinants in its linear response ansatz. On the contrary, for the correct depiction of the environment interactions, one needs to account for polarizability in a balanced manner. These two challenges are successfully addressed by the recently developed hybrid QM/EFP methods, with equation-of-motion coupled-cluster as the QM method of choice.
The result is an efficient method to estimate excitation energy, ionization energy, electron affinity, and redox potential in the condensed phase. It has further been extended to biological systems.
Lower scaling Spin Flip methods
Spin flip equation of motion coupled cluster (EOM-SF-CC) can correctly treat situations involving electronic degeneracies or near degeneracies, e.g., bondbreaking, di- and tri-radicals, etc. However, for large systems EOM-SF-CC (even in single and double excitations) is computationally prohibitively expensive. We develop lower computationally scaling variants of the method.
Density Matrix Renormalization Group
It can be shown that the density matrix renormalization group (DMRG) enables near-exact calculations in active spaces much larger than are possible with traditional complete active space algorithms. We have implemented orbital optimization with the DMRG to further allow the self-consistent improvement of the active orbitals, as is done in the complete active space self-consistent field (CASSCF) method.
Machine learning based force fields
Polarizability is a crucial component for the force fields used in hybrid QM/MM methods for excited state properties. However, the polarizable force fields are in most cases computationally expensive. We develop polarizable force fields based on supervised machine learning to be as accurate as QM methods and considerably cheaper computationally.