Physics-based, data-driven cell-scale membrane simulations with HMFF

Le résumé fourni par la source

This record contains data and analysis code for the manuscript titled Physics-based, data-driven cell-scale membrane simulations with HMFF. The following describes the layout and content of this record. Please note that the directories outlined below have been uploaded as zip archives. analysis_code/ Analysis scripts organized by biological system and analysis type: `hmff.R` Core HMFF analysis and visualization functions. `dts_screens.R` Additional screens and validation experiments. `synthetic/create_map.py` Generate the synthetic planar Gaussian target density. `synthetic/dts_to_density.py` Utility to obtain a simulated membrane density from a mesh. `synthetic/analyze_runs.py` Utility to analyze DTS validation runs. `synthetic/dts_undulation.py` Utility to analyze mesh undulations. `synthetic/analyze_undulation.py` Compute undulation spectra using `dts_undulation.py`. `iav/density_line.py` M1 density profile along the viral envelope cross-section. `iav/project_picks.py` Project HA/NA picks onto the mesh, compute curvature and geodesic distances. `mito/mito_stats.py` Triangle count, vertex count and area per mitochondrion. `mito/mito_features.py` Feature set per mitochondrion (area, volume, segments, curvature, eigenvalues). `mito/zernike.py` Rotation-invariant 3D Zernike descriptors. `mito/mito_clustering.R` t-SNE embedding and Louvain clustering of the feature set. `mito/make_sessions.py` Split the mitochondria session by Louvain cluster. `mito/mito_contacts_ray.py` Ray-cast contacts between mitochondria and surrounding organelles. `mito/merge_mito_contacts.R` Threshold and aggregate contact area per mitochondrion. `mito/mito_hmff_stats.py` Per-frame mesh properties along the mitochondria HMFF trajectory. data/ Organized by system, one directory per system. Each contains a mixture of inputs (densitymaps, DTS input decks) and the analysis outputs derived from them; the tables listed here arewhat `analysis_code/hmff.R` and `analysis_code/dts_screens.R` read to produce the figures.Where a directory contains a `dts/` subdirectory, which holds the FreeDTS input deck(`input.dts`, `topol.q`, `top.top`, `run.sh`) needed to repeat the corresponding simulation. Synthetic validation systems `plane/` - Planar membrane lambda screen. `plane_potential.txt` is the synthetic Gaussian potential; `plane_hmff_{0,05,1,2,5,10}.csv` are the recovered potential profiles at lambda = 0, 0.5, 1, 2, 5, 10; `VTU_F_MDFF_*` hold the corresponding FreeDTS trajectory frames (`.vtu`). `dts/` includes the target `map.mrc`. `undulation/` - Undulation spectrum runs `run_0001` ... `run_0011`, each with its `input.dts`, `params.json` and the `all_/mean_/std_spectra.csv` computed by `analysis_code/synthetic/analyze_undulation.py`. `map.mrc` is the target density, `screen.dts` / `topol.*` the screen definition. `bud/` - Budding membrane with curvature-inducing inclusions. Per-vertex `mean_curvature_vertex.txt`, `gaus_curvature_vertex.txt`, `bending_energy_vertices.txt` and `hmff_potential_vertices.txt` from the HMFF run, alongside the ground truth the target was generated from (`ref_mean_curvature_vertex.txt`, `ref_inclusions.txt`). `dumbbell/` - Screen driving a vesicle to a dumbbell shape, three replicates (`dumbbell_{0,1,2}.txt.gz`). `discoid/` - Same for the discoid shape (`discoid_{0,1,2}.txt.gz`), plus two larger parameter screens, `screen_discoidXiTotal.txt.gz` (over xi) and `screen_discoidXiTotalLambda.txt.gz` (over xi and lambda jointly). M. pneumoniae' `mycoplasma/` `mycoplasma_00242_6.80Apx.mrc` - cryo-ET tomogram. This tomogram was collected for Xue et al., Nature (2022); as it has not been deposited elsewhere, we include it here. `mean_curv.csv`, `curvature_data.csv`, `edge_lengths.csv`, `area_comparison.csv`, `volume_comparison.csv` - mesh geometry before/after HMFF. `ribo_distance_all.csv` - ribosome-to-membrane distances. `00242_density.csv` - input density profile. `00361.csv`, `00423.csv`, `00429.csv`, `00361_boxed.csv` - convergence traces of the validation runs in `mycoplasma_validation/`; `rmap.txt` maps each `run_N` in `00361_boxed.csv` to the target map it was fitted against `mycoplasma_validation/` - `dts/{00361,00423,00429}/`, the FreeDTS decks and meshes (`mesh_base.q`, `mesh_equilibrated.q`, `mesh_remeshed.q`, `screen.dts`) for the three validation cells. The input tomograms and membrane segmentations for these three cells are from Siggel et al., J. Struct. Biol. (2024) and are not redistributed here. Influenza A virus (iav/) `emd_11075_smoothed.mrc` is the filtered EMD-11075 map used for HMFF. `ha.tsv` / `na.tsv` are the template-matching picks for HA and NA. `ha_geodesic.csv` / `na_geodesic.csv` the geodesic nearest-neighbour distances between them. `dist_*.csv` and `angle_*.csv` compare glycoprotein placement across the HMFF, HMFF+lambda, raw segmentation and cylinder models. `projection_curvature.csv` and `protein_curvatures.csv` give per-protein membrane curvature. `density_line.csv` and `iav_density_curve.xyz` the M1 density profile. `seedpoint_distances.csv` / `seedpoint_mesh_distances.csv` the seed-point accuracy check. HeLa cell / mitochondria (mitochondria/) `mitochondria/` - All derived from `mosaic_sessions/mito_remesh.pickle` via the scripts in `analysis_code/mito/`. `features/` - morphological feature tables per mitochondrion (`features_combined.csv` and its z-scored counterpart, `area.txt`, `volume.txt`, `segments.txt`, `eigenvalues.txt`, `mean_/gauss_curvature*.txt`, `zernike/`), the Louvain assignment used in the paper (`clustering.txt`), an independent run using recommended tsne parameters from Kobak and Berens (`clusteringSnifter.txt`), and `mito_stats.csv` `mito_ray_distances.csv`, `mito_ribo_contacts_ray.csv`, `ribo_dist_ray_filter.csv` are ray-cast mitochondria-ER contacts and ribosome occupancy. `dts_en.csv`, `distances.csv`, `mito_trajectory.csv` - HMFF parameter screen over (kappa, xi, lambda); run ids encode the parameter set, e.g. `mito_20_15_50`. `mito_20_15_50/` - full DTS trajectory (101 `.tsi` frames) for the parameter set used in the figures. `sessions/` - one Mosaic session per Louvain cluster (`1`-`9`), plus `total.pickle` and `scales.txt`. Regenerable from `mito_remesh.pickle` with `analysis_code/mito/make_sessions.py`. External datasets (publicly available, not included here): Influenza A virus tomogram: EMD-11075 (EMDB) HeLa cell FIB-SEM data: OpenOrganelle dataset jrc_hela-2 (https://open.quiltdata.com/b/janelia-cosem-datasets/tree/jrc_hela-2/); see `source_code/download.py` M. pneumoniae ribosome template: EMD-17132 (EMDB) M. pneumoniae Nap template: EMD-17591 (EMDB), atomic model PDB 8PBZ M. pneumoniae validation tomograms and segmentations: Siggel et al., J. Struct. Biol. (2024) mosaic_sessions/ Complete Mosaic session files containing the processed geometry for each system. Open them from the GUI or from Python via `mosaic.session.open_session()`. `cell_full.pickle` All segmented organelles of the OpenOrganelle jrc_hela-2 volume, meshed as described in the methods. Superset of `cell_hmff.pickle`. `cell_hmff.pickle` HMFF HeLa organelle surfaces. `mito_remesh.pickle` Remeshed mitochondrial surfaces extracted from the jrc_hela-2 session. This is the input for every script in `analysis_code/mito/` (`make_sessions.py`, `mito_features.py`, `mito_stats.py`, `zernike.py`, `mito_contacts_ray.py`). `iav_full.pickle` Influenza A virus session (EMD-11075) with membrane segmentation and template-matching picks. Input to `analysis_code/iav/project_picks.py`. `iav_hmff.pickle` HMFF influenza envelope. `mycoplasma.pickle` M. pneumoniae cell session (tomogram 00242), including the HMFF membrane and ribosome positions. `plane_hmff.pickle` Planar membrane validation session covering the lambda screen in `data/plane/`. worked_example/ Self-contained example for reproducing the influenza virus analysis (see separate README.txt within this directory). This also includes HA/NA structure predictions used for template matching and the coarse

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.