← Back to Blog

What Is Multi-Ligand Simultaneous Docking? A Guide to How Combination Drug Discovery Actually Works

September 2026 · Ashwin Desai

Most molecular docking studies ask a simple question: how well does this one molecule fit into this one protein pocket? That question has driven decades of computational drug discovery, but it misses something increasingly important in modern medicine - most effective therapies aren't single molecules acting alone. They're combinations, and combinations can behave very differently from the sum of their parts.

Multi-ligand simultaneous docking (MLSD) is a computational technique built to answer the harder question: what happens when two or more compounds occupy a binding pocket at the same time? Across four of our studies - spanning cancer targets and, more recently, Alzheimer's disease - the answer has consistently been that combinations can bind more tightly than any single compound, including established commercial drugs.

How MLSD differs from standard docking

Conventional molecular docking scores one ligand against one target and produces one binding affinity, typically expressed in kcal/mol (a more negative number means tighter, more favorable binding). This is useful, but it treats every candidate compound as if it will act in isolation - which is rarely how real pharmacology works, especially for compounds derived from natural plant extracts that contain dozens of active phytochemicals simultaneously.

MLSD instead docks multiple ligands into the same pocket at once and evaluates the resulting complex as a system. This surfaces three distinct patterns:

• Additive effects - ligands bind at distinct, non-overlapping regions within the same pocket, each contributing independently to overall stability.
• Allosteric effects - one ligand binding at a secondary site changes the protein's conformation in a way that improves the other ligand's binding at the primary site.
• Synergistic effects - the combined binding affinity substantially exceeds what either ligand achieves alone, sometimes even exceeding commercial single-agent drugs.

Case 1: Withaferin A and Garcinol against BCL-2 and AKT-1

In our first application of MLSD, we tested Withaferin A (from Ashwagandha) and Garcinol (from Kokum) - two phytochemicals long used in traditional medicine - against two cancer-associated proteins. Docked together, they reached a BCL-2 binding affinity of −11.88 ± 0.12 kcal/mol, exceeding the commercial inhibitor venetoclax (−9.73 ± 0.1 kcal/mol). Against AKT-1, the combination reached −13.74 ± 0.08 kcal/mol - nearly double the affinity of melatonin (−7.24 ± 0.06 kcal/mol), a known AKT-1 inhibitor.

Read the full study → (DOI)

Case 2: Moringa oleifera phytochemicals against BCL-2

Moringa oleifera leaf extract contains dozens of bioactive compounds. Using MLSD, we found that apigenin, hesperetin, and niazimicin A bound BCL-2 together with an affinity of −14.96 kcal/mol - again surpassing venetoclax. This was the first application of MLSD specifically to a plant extract's full phytochemical profile, rather than two isolated compounds.

Read the full study → (DOI)

Case 3: Carpaine and Rutin against BCL-2 and WWP1 - two mechanisms at once

This study, published in Phytomedicine Plus, is where MLSD revealed something more nuanced: the same two compounds acted through different mechanisms depending on the target. Against BCL-2, Carpaine and Rutin showed an additive effect, binding at distinct regions of the same pocket to reach −13.13 ± 0.08 kcal/mol, again beating venetoclax. Against WWP1, the effect was allosteric - Carpaine bound near one site while Rutin bound near another, and their interaction increased Carpaine's effective binding affinity to −15.59 ± 0.39 kcal/mol. When Rutin was further combined with the clinical drug bortezomib, the interaction with WWP1 became more strongly synergistic still, suggesting a plant-compound-plus-clinical-drug combination strategy.

Read the full study → (DOI · arXiv preprint)

Case 4: A first application to Alzheimer's disease

Every prior MLSD study we'd run was in oncology. In 2025, we applied the same method to a very different problem: Alzheimer's disease, where no disease-halting therapy currently exists and BACE1-targeting drugs have struggled with side effects and insufficient efficacy. We screened 15,641 small molecules with known activity against BACE1, identifying five with strong individual binding affinities (−11 kcal/mol or better). Four combinations of these - including CHEMBL4078427 paired with CHEMBL3656158 and the clinical candidate lanabecestat - reached MLSD binding affinities of −19.90 to −17.67 kcal/mol, far exceeding any single-ligand result. Molecular dynamics simulations confirmed the combined complexes were more stable than single-ligand binding, supporting combination therapy as a viable strategy against BACE1.

Read the full study → (DOI · arXiv preprint)

How MLSD actually runs: it's simpler than it sounds

Despite the results above, MLSD isn't a separate, exotic piece of software - it's the same docking engine used for standard single-ligand docking, run with one small change to the command. Using AutoDock Vina, the two commands below (from our Carpaine/Rutin study) differ only in the --ligand argument: single docking takes one ligand file, MLSD takes multiple.

Single-ligand docking

vina_1.2.5_linux_x86_64 --receptor receptor.pdbqt --ligand ligand.pdbqt --config config.txt --exhaustiveness 32 --out output.pdbqt --num_modes 3 --verbosity 2

Multi-ligand simultaneous docking

vina_1.2.5_linux_x86_64 --receptor receptor.pdbqt --ligand ligand1.pdbqt ligand2.pdbqt --config config.txt --exhaustiveness 32 --out output.pdbqt --num_modes 3 --verbosity 2

Every other parameter - exhaustiveness, number of output modes, verbosity - stays identical. Vina evaluates the additional ligand file as part of the same simultaneous search rather than in a separate run, which is what allows the resulting pose to capture ligand-ligand interactions (additive, allosteric, or synergistic) rather than two independent, non-interacting binding events. The barrier to running MLSD yourself, in other words, is conceptual rather than technical - the harder part is deciding which ligand combinations are worth testing in the first place.

Read the full Carpaine/Rutin study → (DOI)

What this means for drug discovery

Across four independent MLSD studies and four different disease targets, one pattern holds: combinations of compounds - whether two plant-derived phytochemicals, or a natural compound paired with a clinical drug - can consistently outperform single-agent binding affinity, sometimes by a wide margin. As computational screening pipelines mature, MLSD offers a way to identify these combination effects computationally before committing to the far more expensive process of experimental combination testing, narrowing the search space for synergistic drug pairs across both oncology and, as our most recent work shows, neurodegenerative disease.

If you're working on combination therapy strategies, multi-target inhibitors, or phytochemical drug discovery and want to discuss MLSD methodology, get in touch.