Use Lacuna Online

Lacuna: Cryptic Binding Pocket Discovery from Conformational Ensembles

Lacuna is an open-source Python tool for discovering cryptic binding pockets: sites that are absent or too small to detect in a protein's unbound structure and open only during conformational fluctuation. It was developed by Clayton W. Moore (Texas A&M University) and is released under the MIT licence.

Many disease-relevant proteins are called undruggable because their experimental structures show no well-formed pocket. Most binding-site predictors score a single static structure, which is exactly the structure in which a cryptic site is invisible. Molecular dynamics can expose the missing conformations, but it places a simulation budget between the user and an answer.

Lacuna takes a cheaper route. It generates a conformational ensemble from any input structure, detects pockets in every conformer, clusters the detections into persistent sites and ranks those sites with a fitted model.

How Lacuna Finds Cryptic Pockets in Protein Conformational Ensembles

1. Generate an ensemble: by default, normal mode analysis produces plausible conformers in seconds with no simulation setup, force field parameterisation or GPU.

2. Detect pockets per conformer: a geometric detector runs independently on every conformer.

3. Cluster into sites: transient detections are matched across conformers into persistent sites that carry ensemble statistics, such as how often the site is open and how much it opens relative to the input.

4. Rank sites: a fitted linear model orders the sites so likely cryptic pockets appear near the top.

5. Export for docking: each site is written as a Boltz constraint, a Vina box and a pseudoatom PDB.

Lacuna Benchmark Results for Cryptic Site Recovery

The paper reports the share of cryptic sites recovered in the top five predictions.

Ranking and Baseline Comparisons

• The fitted ranker recovers 55.6% of CryptoBench test-fold sites, against 17.8% for the earlier analytic crypticity rule.

• With the identical normal-mode ensemble, MDpocket recovers 43.9%, so the gain comes from clustering and ranking.

• P2Rank is better on general holo sites (93.8% against 86.8% on 144 paired structures). On cryptic sites the PLM-assisted ranker reaches statistical parity with P2Rank, while the default trails it.

In a worked example on apo K-Ras (PDB 4OBE), the switch-II pocket is ranked first, found in 11 of 20 generated conformers and in none of the input structure, growing from 0 to 305 ų.

What is Tamarind Bio?

Tamarind Bio is a no-code bioinformatics platform built to give life scientists and researchers access to powerful computational tools. Many cutting-edge machine learning models are hard to deploy and use. Tamarind provides an intuitive, web-based environment that removes the complexity of high-performance computing, software dependencies and command-line interfaces.

The platform is designed for biologists, chemists and other researchers who may not have a background in programming or cloud infrastructure but want to run models on their own data. Key features include:

• A user-friendly graphical interface for setting up and launching experiments

• A robust API for integration into existing research pipelines

• An automated system for managing and scaling computational resources

Tamarind treats information and data security as a top priority, as detailed in its Trust Center and Terms of Service.

Accelerating Discovery with Lacuna on Tamarind Bio

• Targeting undruggable proteins: look for hidden pockets on proteins with no obvious binding site in the apo structure.

• Structure-based virtual screening: hand ranked sites directly to docking through Vina boxes or Boltz constraints.

• Prioritising sites: use the crypticity score and site-opening statistics to decide which pockets deserve follow-up.

• Comparing sampling methods: swap in MD, Boltz-2 or your own ensemble without changing downstream analysis.

How to Use Lacuna on Tamarind Bio

1. Open the tool: log in to tamarind.bio and select Lacuna.

2. Provide a structure: upload an apo protein structure (PDB), the input the method is built around.

3. Choose the ensemble backend: keep the default normal mode analysis, or pick another backend or supply your own ensemble if exposed.

4. Set parameters: adjust options such as the number of conformers or the ranker if exposed.

5. Run the job: the default backend takes seconds per chain.

6. Download outputs: collect ranked sites with crypticity scores, Boltz YAML constraints, Vina boxes and pseudoatom PDB files.

7. Use downstream: dock ligands into the top sites or validate them with simulation.

Things to Keep in Mind

• Lacuna targets cryptic sites, not general binding-site prediction; P2Rank is better on already-open holo sites.

• The correct site is often found but out-ranked: some candidate clears the criterion for 73.7% of test-fold structures, against 66.1% in the top five.

• The default elastic network is harmonic and cannot generate large hinge or interface openings; Lacuna recovers 47% of the most-mobile quartile against 62% for P2Rank.

• The default ranker does not reach parity with P2Rank on cryptic sites; the paper is a preprint and has not been peer reviewed.

Source: Moore CW. Lacuna: Cryptic Binding Pocket Discovery via Conformational Ensemble Analysis. bioRxiv preprint, 2026. DOI: 10.64898/2026.08.14.744956. Code: https://github.com/mooreneural/lacuna.