Pauling vs. AutoDock Vina

Pauling runs UniDock -- a GPU-accelerated fork of Vina with the same scoring function -- at 1000x throughput with automatic receptor prep, box placement, and format conversion. Zero setup. Vina itself is the most-used docking engine in academia -- free, reliable, well-cited -- but you're on your own for prep, conversion, and scaling beyond one machine.

Category Pauling AutoDock Vina
Docking engineUniDock -- GPU-accelerated Vina fork, same scoring functionOriginal Vina -- most-cited docking engine, extensively validated
GPU accelerationUniDock runs on GPU -- massively parallelCPU only; one ligand at a time per core
Throughput10M+ compounds on auto-scaling Cloud DataflowHundreds to low thousands per day on a single workstation
Receptor preparationAutomatic -- protonation, format conversion handled by PaulingManual -- MGLTools/ADFR Suite for PDBQT conversion, charge assignment
Box placementAutomatic via P2Rank pocket detection or user-specifiedManual -- you define center and dimensions yourself
File format handlingAutomatic conversion (PDB, SDF, MOL2, SMILES, CIF, PDBQT)PDBQT only -- you handle every conversion
ScalingAuto-scales to hundreds of workers; fault-tolerant with retriesDIY -- write your own parallelization scripts or use GNU Parallel
ADMET integrationRun ADMET profiling in the same workflow as dockingNone -- separate tool, separate pipeline
Pose quality controlPoseBusters automatic QC on all docked posesNone -- manual inspection of results
MD follow-upGROMACS MD with auto parameterization, MM-PBSA -- one conversationSeparate setup: GROMACS/AMBER install, topology prep, run scripts
CostPay-per-compute on cloud infrastructureFree and open-source
Scoring functionVina scoring (via UniDock) -- same empirical functionVina scoring -- identical, battle-tested
ReproducibilityFull job logging -- every parameter, input, output trackedDepends on your own record-keeping