PharmCast™

A pharmacophore fingerprint records the three-dimensional arrangement of binding features a molecule can present, so two compounds from completely different scaffolds can be compared on the thing a protein actually reads. That encoding is the 10,560-bit three-point PharmPrint™, and comparing two of them is PharmSim™.

Its cost has never been the fingerprint. Of the 2.48 seconds needed to fingerprint one catalogue molecule over 100 conformers, 2.44 s is conformer generation and only 0.043 s is the fingerprint itself. PharmCast is a model that predicts the complete fingerprint directly from a SMILES string, removing the conformational stage rather than accelerating it. That is why the speed-up is of a different order to what optimisation normally buys.

The fingerprint

A fixed geometric vocabulary, not a substructural one

Every pharmacophore in the scheme is a triangle of three typed features at three measured distances. Enumerate every combination of feature types and distance ranges and you get a fixed vocabulary of possible triangles. Each one is a bit. A molecule sets the bit for every triangle it can actually form.

Because the vocabulary is geometric rather than substructural, two molecules with nothing in common on paper can score highly against each other if they present their features in the same places. That is exactly the comparison a scaffold hop needs, and it is not what a two-dimensional fingerprint measures.

What PharmCast predicts

PharmCast reads a canonical SMILES and predicts all 10,560 bits, one output per bit. It was trained on ensemble fingerprints, the bitwise OR of 100 conformers per molecule, so what it reproduces is the ensemble fingerprint and the similarity between two of them. The output is the standard native format: the same 330 unsigned 32-bit words the real calculation emits, in the same bit convention, so existing PharmSim tooling consumes it unchanged.

Accuracy

Measured on molecules the model has never seen

That last row is the one that sets the scale. Rebuilding the same molecules with a different embedding seed reproduces pair similarity to 0.006 at r 0.995, so the reference calculation agrees with itself an order of magnitude more tightly than PharmCast agrees with it. The ground truth is not the problem, and 0.006 is the floor no surrogate can beat.

Where it holds, and where it does not

The reason is entirely in the training corpus, and it is worth stating plainly. The screening collection is filtered at 600 Da on ingest, so it contributes nothing above that line by construction. Only about 37,060 training molecules, 1.43% of the corpus, sit above 600 Da, and every one of them is a peptide. Above 800 Da it is 0.10%. So when PharmCast is asked about a large non-peptide drug it is extrapolating from a few thousand peptides, and its error rises from 0.02 to 0.07 while r falls from 0.97 to 0.63. This is a genuine size effect and not simply a harder task: a matched-spread catalogue control gives 92% on the identical protocol.

What it is trained on

Every molecule fingerprinted with the real calculation, not a shortcut

The screening collection

The source is the June 2026 release of the Enamine screening collection, which currently states 4,774,670 compounds. These are real, orderable material: compounds synthesised and held in stock, quality controlled to at least 90% purity, not a virtual enumeration. Enamine's much larger combinatorial spaces are not included.

Property filters on ingest keep compounds up to 600 Da, at most 10 rotatable bonds and no more than one Lipinski violation; 4,612,044 survive. That filtered set is the index PharmCast draws from, and it is being fingerprinted continuously with the real 100-conformer calculation. SP v4 was trained on the 2,511,440 finished at the time of its snapshot. The 600 Da filter is applied here, on ingest, which is precisely why the model has no catalogue chemistry above that line to learn from.

This is a lead-like library rather than a fragment or biologics-adjacent one. The median compound weighs 344, carries three rings of which two are aromatic, five rotatable bonds, and sits at cLogP 2.7. Only 8% carry defined stereochemistry, so the collection is predominantly flat, achiral scaffolding decorated with substituents.

Wide in frameworks, deep in analogs

Two measurements of the same collection pull in opposite directions, and both are true. Reduced to Bemis-Murcko scaffolds it is broad: 679 distinct frameworks per thousand molecules, 88% of them appearing exactly once, and the ten commonest together account for only 4.0%. But compared whole-molecule against whole-molecule (exhaustively, all 4,612,044 against each other with no sampling), it is dense.

The resolution is that the collection is wide in frameworks and deep in analogs around them: many scaffolds, each decorated many times over with closely related substituents. That has a direct consequence for how any model on this data must be evaluated: a random split places near-identical analogs on both sides and reports an optimistic number.

The peptide corpus, and why it is not just more molecules

An RCSB query at 2.5 Å resolution or better with R-free at or below 0.22 returned 71,018 entries, of which 69,450 yielded usable fragments: 1,524,535 loop instances across 135,408 chains, collapsing to 117,741 distinct sequences and 133,136 distinct capped molecules. Loops are runs of two to six residues lying between annotated secondary structure, read from the helix and strand records in the mmCIF itself rather than re-derived.

Each fragment is capped with its real flanking atoms (an acetyl built from the preceding residue, an N-methylamide from the following one), so it never presents a free amine and acid the protein does not have. Backbone continuity is enforced by requiring the carbon-to-nitrogen distance between consecutive residues to fall between 1.0 and 2.0 Å rather than trusting residue numbering, which rejected 54,520 candidates that numbering alone would have accepted.

Ensemble enhancement is the novel step. Each distinct peptide gets one fingerprint that ORs together every rigid crystallographic conformation actually observed for it in the PDB with 100 computed conformers, generated exactly as for the screening collection. The conformer count therefore runs above 100, and the fingerprint carries both what nature was caught doing and what the molecule can do.

Repeats are kept rather than deduplicated, and the reason is measurable. Across 1,500 sequences seen in three or more entries, the median agreement between observations of the same sequence is 0.878, but 18% fall below 0.70, carrying genuinely different pharmacophores from one crystal structure to the next, and only 19% are near copies. Collapsing them to one representative would throw that away.

Model card

In progress not yet released

A corpus of real, activity-backed ChEMBL molecules in the 600–1000 Da band is being fingerprinted to close exactly the gap described above; 134,765 molecules carry a measured pChEMBL value and qualify. The model that includes them will be PharmCast SCP: Screening collection, ChEMBL, Peptides. No results are claimed for it here, and none should be assumed until it is measured on the same held-out set.

Using it

The Python package is small and self-contained. words_batch is the API that matters: the model earns its speed in a batch, and a single call is dominated by featurisation overhead.

from pharmcast import PharmCast, pharmtan, pharmsim

pc = PharmCast.load("pharmcast_sp_v4.pt")
a, b = pc.words_batch(["CCO", "c1ccccc1O"])   # 330 uint32 each

pharmtan(a, b)          # the PharmSim coefficient
pharmsim(a, b)          # ...with the shared and exclusive bit sets shown

Utilities for going from 2D structure to a native fingerprint, for PharmSim / PharmTan comparison, and for expanding a fingerprint into its 0/1 bit sets will be released as an open-source package. The reference fingerprint generator is not part of that release.

Citation