Genomics Engine
A decentralized, open, AI-driven ecosystem where complex biological data is transformed into actionable human insight. By merging knowledge gained by scientific perseverance with an emergent era of LLMs, Multi-Agents workflows, and a "hardcore" ethical framework, we are ending the era of academic research in silos and returning the code of life ready to be explored by individuals from any field.
In decoding life, everyone is invited to contribute.
What if all human genes can tell their stories? ..all 62,700 of them !
Imagine if every single gene, pseudogene, lncRNA, all functional regions across all chromosomes discovered so far - can tell their story, their contribution in making you, you.
Now, you can read these genes like a book.
Application Features:
Book-like reading interface with chromosome as chapters
Chromosome ideogram showing exact gene location with G-band colouring
Gene stories collated by AI Claude Sonnet 4.6 model
Information based on academic research for each gene annotation.
Real-time streaming for new stories
Gene search across all chromosomes
Gene-Intel is a novel computational platform that reframes genomic search as a structural and spatial reasoning problem rather than a sequence alignment problem. By encoding gene architecture (exon–intron structure, coding sequence composition, UTR profiles) and chromosomal neighbourhood context into a property graph database, and coupling it with a natural-language AI interface, Gene-Intel enables researchers, clinicians, and students to discover functional gene twins - genes that perform equivalent biological roles across wildly divergent species in simple conversation like interface.
Application Features:
The current MVP indexes 15 species spanning five kingdoms of life (Animalia, Plantae, Fungi, Bacteria, and Algae)
Ingests ~1.4 GB of compressed annotation data, extended on regular basis.
Serves interactive WebGL graph visualisations of up to 300 genes alongside persona-aware AI explanations.
Map your gene's address in 3D space. Discover what makes it druggable.The human genome is not a flat linear array of gene sequences - it folds into an intricate 3D architecture inside the nucleus (like a bowl of ramen soup). The location of a gene in this 3D space determines which enhancers switch it on, which proteins it interacts with, how "posh" the locality is and whether it can be safely targeted by a drug or a gene editor.Gene-Maps translates this 3D genome science into an accessible research tool, like exploring the place on a map before you visit. Search any human gene and instantly explore its spatial context: how it interacts with neighbours in chromatin space, how conserved its sequence is across evolution, how risky a CRISPR edit would be at a given position, and how druggable it looks from a spatial genomics perspective.First platform to provide an early scouting route to save millions of $$ in further research.
Application Features:Composite 3D Map Spatial Score : Aggregates five independent data dimensions (evolutionary conservation, chromatin accessibility, PPI network centrality, Hi-C contact frequency, and GTEx tissue expression) into a single weighted druggability index, replacing manual lookups across five separate databases.CRISPR TAD Disruption Risk : Calculates edit safety by querying UCSC, ENCODE, CTCF occupancy in a ±50 kb window and PhyloP100way constraint at the cut site, quantifying proximity to Topologically Associating Domain boundaries — a spatial off-target mechanism that sequence-only tools (CRISPOR, Cas-OFFinder) miss entirely.Evolutionary Distance-Weighted Conservation : Queries ENSEMBL REST for orthologs across 10 model organisms (mouse → nematode, spanning 900 Mya of divergence) and weights each species by phylogenetic distance to produce a biologically interpreted per-gene conservation card.Fully Deterministic, Reproducible Scoring : All five scoring modules use real API calls or tier-based heuristics with no random seeds or stochastic components; identical inputs always return identical outputs, a hard requirement for reproducible scientific research.Multi-tier Live Data Pipeline : Routes each query through ENSEMBL, UCSC, STRING DB, and GTEx APIs with a Redis/Upstash cache layer; Neo4j AuraDB stores the Hi-C contact graph and PostgreSQL holds gene metadata, keeping computation and storage concerns cleanly separated.
Decentralized research on Bio-safety, AI, overlapping ethics and principles of genomics research + AI community on holistic development of human well being.