Last updated: 2024-12-18, by Jacob Williams

These are supplemental data for "Computed isotope shifts of high-frequency vibrational modes exceed thermal noise in propionate bound to a human olfactory receptor", by Jacob Z. Williams, Piotr E. Marszalek, and Weitao Yang. The associated manuscript is available [on the ChemRxiv preprint server](https://doi.org/10.26434/chemrxiv-2024-242tq). The project was funded by the U.S. National Science Foundation (MCB-2105612) and National Institutes of Health (5R01GM061870-20).

The protein/ligand structure and MD trajectory can be downloaded [from the GPCRmd repository, record 1245](https://www.gpcrmd.org/dynadb/dynamics/id/1245/).

# Data files

Each data directory is compressed in an archive, `data-*.tar.gz`.
`data-ash/`: ASH input and output files, as well as intermediate steps for  generating the ligand topology.
* Includes the `mamba` (Python + packages) environment as `ASH-new.yaml` as well as another `README` files for more details.
* Preparation of the ligand topology takes place in this directory, using `ligand_topol.sh`, `sort_mol2_bonds.pl`, and `protein_topology.sh`. (The `topology.md` file provides more details.)
* CHARMM36m force field files are in `charmm36-jul22.ff/`.
* ASH input/output files are in `isotopes-all/NNNns/` (for the snapshot at `NNN` ns). The main output files are `ash_input.out` and `out/vib_ISOM.csv` (these are renamed `NNNns_ISOM.csv` and copied to `data-plot/` for visualization).
`data-basis/`: ASH run on one snapshot (see above) with DFT computed with both triple- and quadruple-zeta basis sets (aug-cc-pVTZ and aug-cc-pVQZ, in `tz/` and `qz/` respectively). Demonstrated that triple-zeta was sufficient.
`data-plot/`: The vibrational spectra and the scripts that plot them:
* `NNNns_ISOM.csv`: The vibrational spectra from the `NNN`-nanosecond snapshot, with isotopic composition given by `ISOM` (e.g. `hhdhd`)
* `avg-ir.csv`: The spectra for each isotopologue averaged over time. Generated by the R script `vib_average.R`.
* `vib_average.R`: An R script that reads in all `NNNns_ISOM.csv` and generates `vib_average.csv`. Requires packages `tidyverse` and `docstring`.
* `plot_avg_ir.py`: A Python script that reads `vib_average.csv` and plots each isotopologue's averaged IR spectrum, as `vib_ISOM.pdf`, as well as a spectrum `vib_all-h-or-d.pdf` that overlays the H$_5$ and D$_5$ isotopologues. Requires libraries `os`, `numpy`, `pandas`, `plotnine`.
* `vib_ISOM.pdf`, `vib_all-h-or-d.pdf`: See above.
* `gas-phase-propionate/`: Contains files `prop-vib-ISOM.csv`, which are IR modes of each isotopologue of gas-phase propionate (without the protein environment).
* `plot_avg_ir_with-gasphase-mol.py`: A Python script, almost exactly analogous to `plot_avg_ir.py`. It also reads the gas-phase data from `gas-phase-propionate/` and overlays those spectra on their respective topologues, so that the averaged effect of the protein environment can be seen relative to the unbound ligand. The files produced are called `vib_ISOM_withgasphase.pdf` as well as `vib_all-h-or-d_withgasphase.pdf`. (These were added in Dec 2024 to make the figures more informative.)
`data-vib/`: The element-by-element decomposition of the normal modes.
* `vib-composition.txt`: Extracted directly from the ASH output files in `data-ash/*ns/ash_input.out`
* `vib-composition.xlsx`: Average elemental composition over the time snapshots
* `vib-composition.csv`: Portable version of the `.xlsx` (but without the formulas)