Reading biology as data, and data as biology.
Biointelligence Lab works at the intersection of genomics, structural chemistry, and machine learning — building tools and models that turn molecular and sequence data into usable biological insight, led by Dr. Udayakumar Mani at SASTRA Deemed University.
PI: Dr. Udayakumar Mani
Institution: SASTRA Deemed University
Email: uthay@bioinfo.sastra.edu
Focus: AI/ML · Bioinformatics · Healthcare
Biointelligence Lab is a computational research group applying artificial intelligence and machine learning to problems in genomics, structural chemistry, and sustainable bioprocessing. The lab builds databases, analysis pipelines, and predictive models that help researchers move from raw molecular and sequence data to actionable biological understanding.
Dr. Udayakumar Mani leads the group's work across two connected fronts: bioinformatics tool-building — metabolome and structural databases, genomic signal processing, and drug-target discovery pipelines — and applied machine learning for chemistry and process engineering, including deep eutectic solvent design, XGBoost-optimised bioprocess extraction, and MLOps compliance frameworks for regulated pipelines.
The group's output spans peer-reviewed work in journals including PLoS One, Future Generation Computer Systems, Journal of Molecular Structure, Scientific Reports, and Separation and Purification Technology, with an author position placing in the top 18 of 32 ranked authors across its impact-factor-indexed journals.
Research Directions
Interdisciplinary work spanning computational biology, machine learning, and structural chemistry.
Structural bioinformatics
Small molecule crystallography, PDB analysis, and residue interaction networks.
AI for healthcare
Deep learning for respiratory analysis, snake species identification, and clinical screening.
Bioinformatics databases
Open-access resources for metabolomics, allosteric regulation, and solvent properties.
Mobile health (mHealth)
Flutter & Android apps for real-time diagnostics and field research.
Research Areas
Five threads, one throughline — each approached with a mix of database engineering, structural analysis, and modern ML.
Genomic & Metabolome Databases
Building and maintaining open resources such as the plant metabolome database (PMDB) and the SWI/SNF remodeling complex infobase for the research community.
Structural & Crystallographic Chemistry
X-ray crystallography, Hirshfeld surface, PIXEL energy, and QTAIM analysis of weak noncovalent interactions in acrylonitrile and pyridine derivatives.
Genomic Signal Processing & ML
Machine learning pipelines for automated detection of cancerous genomic sequences and diseased-gene identification from signal-processed genomic data.
Green Solvents & Sustainable Bioprocessing
Deep eutectic solvent design, ML-optimised extraction (XGBoost, supercritical fluid extraction), and AI-assisted algal biorefinery frameworks.
AIoT, MLOps & Compliance
Applying AI-of-Things process intensification and continuous-compliance MLOps frameworks (AuditOps) to sustainable manufacturing pipelines.
Applied Bio-Tool Development
Desktop and web tools — CASTRA, PDB@, Comat, RSMD, Active Motif Finder — for structural, genomic, and mutation analysis workflows.
Apps & Databases
Field-deployed apps and standalone databases/tools built and maintained for the wider research community.
Applied apps
Safe Breath
Analyses lung sound recordings for preliminary respiratory assessment. Built with Gowri Sankhar S.
- AI-based predictions (e.g. crepitation)
- Built-in guide & disclaimer
Designer Solvent DB
Experimental properties for Deep Eutectic Solvents and Ionic Liquids, filterable by HBA/HBD.
- 500+ DES entries · 50+ properties
- 1000+ peer-reviewed references
Slitherin
Identifies snake species from a photo using deep learning models.
- 14+ species · 4 languages
- Habitat & venom-status info
BrainSpark
Turns abstract learning into interactive neural-style knowledge maps — concepts as neurons, understanding as connection strength — to surface gaps and guide study.
- Visual knowledge mapping
- Instant gap detection & smart study planning
Gene-GPT RAG
Retrieval-augmented Q&A grounded in 35M+ papers and 100+ tool docs, with code generation, conversation memory, and confidence scoring.
- PubMed/MEDLINE, tool docs & community forums
- Research use only — validate experimentally
Embedded Health Monitor
(A) NodeMCU ESP32 (B) MAX30100 pulse oximeter (C) TCS34725 RGB sensor (D) DS18B20 temperature sensor
A health-tech prototype integrating embedded systems, IoT, and real-time monitoring — sensor fusion and scalable data pipelines aimed at making health monitoring more accessible.
Databases & tools
PMDB
1000+ annotated plant metabolites with SMILES, formulae, and structure files, linked to KEGG, PubChem & ChEBI.
Visit PMDB →SWI/SNF Infobase
27 subunit types, 522 non-redundant genes across 20 sequenced organisms.
Visit SWI/SNF Infobase →AlloBase (AlloSDB)
1,457 proteins, 878 allosteric sites across 11 protein families, with drug annotations and signaling networks — connecting scattered allosteric findings into one knowledge network.
Visit AlloSDB →PuckeringDB
Calculates puckering coordinates, classifies ring conformations, and visualises 3D structures & pseudorotation paths, per Cremer & Pople (1975).
Visit PuckeringDB →Non-invasive Vitals Estimation
Estimates heart rate, systolic and diastolic blood pressure from stacked images and derived PPG signals using a combined CNN-LSTM model, five-fold cross-validated.
Try the demo →Active Motif Finder
Finds motif occurrences tolerant of insertions, deletions and mismatches via ASM scoring.
Visit AMF →PDB@
Graphical Viewer, Ramachandran Explorer and PDB Editor, integrating ClustalW.
Visit PDB@ →CIFSYM
Generates unit-cell molecules from a CIF file using the space group's symmetry operations.
Visit CIFSYM →ResComm
Builds residue interaction networks from PDB files; exact Brandes betweenness centrality & community detection.
Visit ResComm →CPIVP
Predicts drug-protein interactions from a protein sequence and compound SMILES using harmonic resonance frequency analysis with knowledge-based enhancement — 100% accuracy, precision & recall on validation test cases.
Visit CPIVP →Impact Metrics
Selected Publications
All tracked publications. Ranked by citation impact — impact factor (IF) shown where indexed.
Contact
Collaborations, students, and data requests welcome.
Reach out about co-authorship, database access, or joining the group.
Dr. Udayakumar Mani
Assistant Professor, Bioinformatics
● Active — recruiting
PI
Dr. Udayakumar Mani