FastSASA

FastSASA calculates solvent-accessible surface area (SASA) for molecular structures and molecular dynamics trajectories. It provides Shrake-Rupley and Lee-Richards calculations through a command-line tool, a Python array API, and VMD integration.

FastSASA is designed for repeated calculations. Trajectory coordinates are streamed in batches while reusable data stays on the selected backend. Vulkan, CUDA, and threaded CPU implementations are included; FastSASA chooses an available backend automatically.

Documentation site · Documentation · CLI reference · Python API · Trajectory guide · VMD integration

Build

Requirements: CMake 3.18 or newer and a C/C++ compiler. GPU support is optional. Vulkan builds also need Vulkan headers and glslc; CUDA builds need the CUDA toolkit.

git clone https://github.com/aminkvh/fastsasa.git
cd fastsasa
cmake -S . -B build
cmake --build build -j4

Install the Python package from the same checkout:

python3 -m pip install .

Platform-specific build options and test commands are in Getting started.

Calculate A Structure

./build/fastsasa --format log structure.pdb
./build/fastsasa --format json --output result.json structure.cif

Shrake-Rupley with 100 points and a 1.4 Å probe is the default. Use --lee-richards for Lee-Richards or --backend cpu|vulkan|cuda to choose a backend explicitly.

Calculate A Trajectory

./build/fastsasa trajectory \
  --topology topology.psf \
  --trajectory trajectory.dcd \
  --frames : \
  --filter protein \
  --output protein_sasa.csv

The default trajectory output is one CSV row per frame. --filter protein defines the atoms included in the calculation; use --select when only a chain, residue, or other subset should be reported in that protein context. See Trajectory analysis and Selection syntax.

Use Python

import numpy as np
from fastsasa import sasa

positions = np.asarray(coordinates, dtype=np.float64)  # (atoms, 3) or (frames, atoms, 3)
radii = np.asarray(atom_radii, dtype=np.float64)       # (atoms,), without probe radius

total_sasa = sasa(positions, radii)

The array API does not require a particular trajectory reader. MDAnalysis, MDTraj, PyMOL, RDKit, feature-extraction, and interface-SASA examples are linked from the Python API.

FastSASA can also turn trajectories into per-residue or named-group exposure time series, interface burial, glycan shielding, summary statistics, and PCA/SVD fingerprints. These outputs support plotting, clustering, rare-state detection, and frame selection. See the Feature Extraction Tutorial for the definitions and worked examples.

License

FastSASA is MIT-licensed. It includes small third-party or attributed data components under their respective licenses; see NOTICE and licenses/.