Molecular properties
Computes the standard small-molecule property profile — molecular weight, LogP, TPSA, hydrogen-bond donors and acceptors, rotatable bonds — from a saved compound structure. Use for a first read on a candidate before ADMET or drug-likeness screening.
Molecular properties
What it does
Takes one saved compound and reports the seven numbers most people mean by "what does this molecule look like on paper":
- Molecular formula and molecular weight (average, the number PubChem itself reports — not the monoisotopic mass spectrometry uses)
- LogP — predicted octanol/water partition coefficient, a proxy for lipophilicity
- TPSA — topological polar surface area, a proxy for membrane permeability
- H-bond donors and H-bond acceptors
- Rotatable bonds — a proxy for conformational flexibility
All seven come back in one run, because nobody asks for molecular weight on its own — these are read together as a first pass on a candidate, before Drug-likeness filters turns them into a verdict or ADMET prediction goes further.
How the numbers are computed
The engine is RDKit, working from the molecule's 2D graph — nothing here needs a 3D conformer. Each property is a standard RDKit descriptor computed directly from the parsed structure: MolWt for weight, MolLogP (Crippen's method) for LogP, TPSA, and simple counts for the rest.
Reading the output
There is nothing to rank or filter — all seven values are shown directly. Take them as estimates from 2D structure, not measurements: LogP in particular is a prediction, and different methods can disagree by half a unit or more on the same molecule.
Input
One compound, saved as an SDF — the format PubChem searches save automatically, or an SDF uploaded directly. If the file has more than one structure in it, only the first is used, and the result says how many records the file had.
Important limitations
2D only. Nothing here reflects a specific 3D conformation. A property that depends on shape rather than connectivity (buried surface area, a docking score) needs a different tool.
No pass/fail judgement. This tool reports numbers, not a verdict — a compound with MW 600 is reported as MW 600, not flagged. Use Drug-likeness filters for the Lipinski/Veber/PAINS read.