Drug Discovery Computational Biology
Computational drug discovery (CADD) uses structure- and ligand-based methods to prioritise compounds before wet-lab testing — molecular docking, virtual screening, QSAR, ADMET filters, molecular dynamics and increasingly AI generative models. Student projects typically combine open tools (AutoDock Vina, RDKit, GROMACS) with public databases (PDB, ChEMBL, ZINC, PubChem).
Below: 80+ topics with tools and representative datasets.
Drug Discovery Computational Chemistry
Tools & PlatformsBest Computational Drug Discovery Project Topics (80+)
Topics with tools and datasets.
| # | Project Topic | Tools | Datasets |
|---|---|---|---|
| Molecular Docking | |||
| 01 | DockProtein–Ligand Docking with AutoDock Vina for a Disease Target | AutoDock Vina · PyMOL | PDB · PubChem ligands |
| 02 | DockRedocking Validation and RMSD Analysis of Co-Crystal Ligands | Vina · RDKit · PyMOL | PDB co-crystal structures |
| 03 | DockFlexible Side-Chain Docking vs Rigid Receptor Comparison | AutoDock · scoring analysis | PDB · known actives |
| 04 | DockMulti-Target Docking of Natural Product Library | Vina · batch scripts · RDKit | ZINC natural · ChEMBL |
| 05 | DockBlind Docking for Binding Site Identification | Vina · grid search · PyMOL | PDB apo structures |
| 06 | DockConsensus Docking: Combining Multiple Scoring Functions | Vina · additional scorers | DUD-E subset |
| 07 | DockCovalent Docking Concepts for Reactive Warheads | specialised protocols · literature | Covalent ligand sets |
| 08 | DockProtein–Protein Interface Docking for PPI Inhibitors Concepts | docking tools · analysis | PPI benchmark complexes |
| 09 | DockDocking-Based SAR Exploration of Analogue Series | Vina · RDKit · plots | ChEMBL analogue series |
| 10 | DockWater-Mediated Interactions in Docked Poses Analysis | PyMOL · hydration analysis | PDB high-res structures |
| Virtual Screening & Hit Identification | |||
| 11 | VSStructure-Based Virtual Screening of ZINC Library vs Target | Vina · RDKit · filtering | ZINC · PDB target |
| 12 | VSLigand-Based Virtual Screening with Similarity Search | RDKit · fingerprints · Tanimoto | ChEMBL actives · decoys |
| 13 | VSHierarchical VS: Pharmacophore Filter then Docking | Pharmit concepts · Vina | DUD-E · ChEMBL |
| 14 | VSEnrichment Factor and ROC Analysis of Virtual Screening | Python · metrics | DUD-E benchmark |
| 15 | VSFragment-Based Virtual Screening and Growing Strategies | RDKit · docking | Fragment libraries |
| 16 | VSNatural Product Virtual Screening against Kinases / Proteases | Vina · NP libraries | ZINC NP · PDB kinases |
| 17 | VSMachine Learning–Boosted Ranking of Docking Hits | scikit-learn · docking scores | Labeled hit/decoy sets |
| 18 | VSMulti-Step Filtering: PAINS, ADMET, then Docking | RDKit · PAINS filters | PubChem · ChEMBL |
| 19 | VSTarget-Focused Library Design from Known Actives | RDKit · scaffold analysis | ChEMBL target actives |
| 20 | VSCross-Docking Study to Assess Pose Prediction Robustness | Vina · multi-structure PDB | PDB conformational set |
| QSAR / QSPR Modeling | |||
| 21 | QSAR2D QSAR Model for Bioactivity Prediction using RDKit Descriptors | RDKit · scikit-learn | ChEMBL bioactivity set |
| 22 | QSAR3D QSAR / CoMFA-Style Concepts with Alignment | alignment tools · regression | Aligned ligand series |
| 23 | QSARQSAR for Toxicity Endpoint Prediction | RDKit · classifiers | Tox21 · related sets |
| 24 | QSARFeature Importance and Interpretable QSAR with SHAP | SHAP · tree models | ChEMBL curated set |
| 25 | QSARScaffold-Split vs Random-Split Validation in QSAR | RDKit · sklearn | ChEMBL series |
| 26 | QSARMulti-Task QSAR for Related Activity Endpoints | multi-output ML · RDKit | Multi-assay ChEMBL |
| 27 | QSARDescriptor Selection and Overfitting Control in Small Datasets | feature selection · CV | Small activity sets |
| 28 | QSARQSPR for Solubility / LogP Prediction Benchmark | RDKit · regression | Public solubility sets |
| 29 | QSARGraph Neural Network QSAR with DeepChem | DeepChem · PyTorch | MoleculeNet benchmarks |
| 30 | QSARApplicability Domain Analysis for QSAR Models | distance metrics · Python | Training space analysis |
| ADMET Prediction & Filtering | |||
| 31 | ADMETIn-Silico ADME Property Prediction Pipeline | RDKit · predictive models | Public ADME datasets |
| 32 | ADMEThERG Liability and Cardiotoxicity Risk Filtering | classifiers · RDKit | hERG assay data |
| 33 | ADMETCYP Inhibition Prediction for Drug–Drug Interaction Risk | ML · descriptors | CYP inhibition sets |
| 34 | ADMETBlood–Brain Barrier Permeability Prediction | RDKit · BBB models | BBB public datasets |
| 35 | ADMETDrug-Likeness and Rule-of-Five / Beyond Rule Filters | RDKit · Lipinski · QED | PubChem / ChEMBL |
| 36 | ADMETPAINS and Aggregator Alerts in Screening Libraries | RDKit · filter catalogs | Screening library sample |
| 37 | ADMETIntegrated ADMET Dashboard for Hit Prioritisation | Python · multi-endpoint scores | Multi-property compound set |
| 38 | ADMETToxicity Endpoint Multi-Label Classification | DeepChem / sklearn | Tox21 |
| Molecular Dynamics of Drug Targets | |||
| 39 | MDProtein–Ligand MD Simulation and Binding Stability Analysis | GROMACS · PyMOL | PDB complex |
| 40 | MDRMSD, RMSF and Hydrogen-Bond Analysis of Docked Complexes | GROMACS analysis tools | MD trajectories |
| 41 | MDFree Energy Concepts: MM-PBSA / MM-GBSA on MD Snapshots | gmx_MMPBSA concepts | MD ensembles |
| 42 | MDLigand Unbinding / Residence Time Insights from MD | enhanced sampling concepts | Long MD trajectories |
| 43 | MDMembrane Protein–Drug Interaction MD Setup | GROMACS · membrane builder | GPCR / ion channel PDB |
| 44 | MDConformational Selection vs Induced Fit Case Study | MD · docking comparison | Apo/holo PDB pairs |
| 45 | MDWater Network Analysis in Active Sites from MD | trajectory tools · PyMOL | MD water density maps |
| 46 | MDForce Field Comparison Impact on Ligand Pose Stability | GROMACS · multiple FF | Same complex multi-FF |
| AI, Generative Models & Deep Learning | |||
| 47 | AIDeep Learning QSAR with Graph Convolutional Networks | DeepChem · MoleculeNet | MoleculeNet datasets |
| 48 | AIGenerative Model for De Novo Molecule Design (VAE / SMILES RNN) | PyTorch · RDKit | ChEMBL SMILES |
| 49 | AIProperty-Optimised Molecule Generation with Reinforcement Learning | RL · RDKit · scoring | Reward-guided design set |
| 50 | AIDrug–Target Interaction Prediction with Deep Models | PyTorch · DTI frameworks | BindingDB · Davis / KIBA |
| 51 | AITransformer Models for Molecular Property Prediction | HF / custom transformers | MoleculeNet |
| 52 | AIActive Learning for Efficient Bioactivity Label Acquisition | modAL · QSAR loop | Unlabeled pool + oracle sim |
| 53 | AIMulti-Modal Fusion: Structure + Ligand Embeddings for DTI | GNN + protein embeddings | DTI benchmarks |
| 54 | AIExplainable AI for Molecular Predictions (Attention / SHAP) | captum · SHAP · RDKit | Trained QSAR / DTI models |
| Pharmacophore & Ligand-Based Design | |||
| 55 | PharmPharmacophore Model Building from Active Ligand Set | Pharmit / RDKit concepts | ChEMBL actives |
| 56 | PharmPharmacophore-Based Virtual Screening and Hit Validation | screening · docking follow-up | ZINC · DUD-E |
| 57 | PharmReceptor-Based Pharmacophore from Docked Complexes | pose analysis · features | PDB complexes |
| 58 | PharmScaffold Hopping Guided by Pharmacophore Constraints | RDKit · scaffold networks | Analogue series |
| 59 | Pharm3D Shape Similarity Screening (ROCS-Style Concepts) | shape tools · RDKit | Query ligands · library |
| 60 | PharmConsensus Pharmacophore from Multiple Crystal Structures | multi-structure analysis | PDB multi-ligand target |
| Target Selection, Pipelines & Capstone | |||
| 61 | AdvDisease Target Prioritisation using Network / Literature Mining | Python · bioinformatics APIs | Open target / literature |
| 62 | AdvHomology Modeling of Target Protein for Docking | Modeller concepts · validation | UniProt · templates |
| 63 | AdvEnd-to-End CADD Pipeline: VS → Dock → ADMET → Rank | Vina · RDKit · scripts | ZINC + PDB target |
| 64 | AdvRepurposing Approved Drugs via Docking against New Target | Vina · DrugBank concepts | Approved drug structures |
| 65 | AdvPolypharmacology Prediction: Multi-Target Ligand Profiles | multi-target docking · ML | ChEMBL multi-target |
| 66 | AdvBinding Site Comparison and Druggability Assessment | fpocket concepts · PyMOL | PDB pocket set |
| 67 | AdvFragment Growing and Linking In Silico Workflow | RDKit · docking cycles | Fragment hits |
| 68 | AdvProspective Virtual Screening Case Study with Retrospective Validation | full protocol · metrics | DUD-E or ChEMBL split |
| 69 | AdvOpen-Source vs Commercial Tool Comparison on Benchmark Set | multiple tools · ROC/EF | DUD-E |
| 70 | AdvReproducible CADD Workflow with Snakemake / Nextflow Concepts | workflow managers · containers | Pipeline config + data |
| 71 | AdvCOVID / Antiviral Target Docking Case Study (Historical Benchmark) | Vina · public structures | PDB viral proteins |
| 72 | AdvKinase Inhibitor Selectivity Profiling by Multi-Kinase Docking | batch docking · heatmaps | Kinase PDB set |
| 73 | AdvNatural Product Priority Ranking with ADMET + Docking Scores | composite scoring · RDKit | NP libraries |
| 74 | AdvInteractive Dashboard for Exploring Docking and QSAR Results | Streamlit · Plotly | Project result tables |
| 75 | AdvBenchmarking Scoring Functions on a Common Pose Set | scoring comparison · RMSD | PDBBind concepts |
| 76 | AdvPrivacy-Preserving / Federated Concepts for Collaborative CADD | architecture discussion | Distributed dataset notes |
| 77 | AdvTeaching Package: From SMILES to Ranked Hits Documentation | full lab manual · scripts | Example target + library |
| 78 | AdvUncertainty Quantification in Docking and QSAR Predictions | ensemble methods · metrics | Calibration plots |
| 79 | AdvMulti-Objective Optimisation: Potency vs ADMET Trade-offs | Pareto analysis · Python | Multi-property compound set |
| 80 | AdvCapstone: Full Computational Hit Discovery Report on Chosen Target | Vina · RDKit · GROMACS optional | PDB + ChEMBL + ZINC |
| 81 | AdvMetabolite Prediction and Soft-Spot Analysis Concepts | metabolism tools · RDKit | Known drug metabolites |
| 82 | AdvOff-Target Prediction and Safety Flagging Pipeline | similarity · panel docking | Off-target panel structures |
Datasets are public (PDB, ChEMBL, ZINC, PubChem, DUD-E, MoleculeNet, BindingDB, Tox21). Always cite sources and respect database licences. Contact us for protocols, pipelines, university-format report, PPT and viva Q&A.
Why Choose Us for Computational Drug Discovery Projects?
Bangalore-based guidance for BE, BTech, MTech and bioinformatics students.
Docking & Virtual Screening
AutoDock Vina pipelines, enrichment analysis and multi-step filtering with open libraries.
QSAR & ADMET
Descriptor-based and deep models, toxicity filters and prioritisation dashboards.
MD & Dynamics
GROMACS setups, stability metrics and free-energy concepts for complex validation.
AI & Generative Design
Graph models, SMILES generators and explainable predictions on public benchmarks.
FAQ — Computational Drug Discovery Projects
Computational Drug Discovery Lab — Bangalore
Docking, QSAR, MD and AI pipeline support for CADD final-year projects.
Docking
Screening
Models
Filtering
Simulation
Design
Screening
Support