01Genome
A genome is a sequence of letters
Your DNA is written in four bases — A, T, G and C — read in order along each chromosome. Nearly all of that sequence is shared by every person alive.
Population-aware genomics
Genomic reference data does not represent India equally.BharatGenome makes that gap visible, and usable.
Search a gene, variant, or rsID. BharatGenome puts Indian population evidence side by side with global reference evidence, explains the difference in plain language, and shows you exactly where each number came from and how far it can be trusted.
India has over 1.4 billion people. The reference resources modern genomics leans on were historically built from a limited set of populations — and allele frequencies differ between populations.
The story
From a sequence of letters to a representation problem
Five steps, no prior genomics needed. Each one is a plain-language idea that the next one builds on.
- 02Variation
The interesting part is where we differ
At some positions, different people carry different letters. Each version is an allele, and how common each allele is in a group of people is its allele frequency.
- 03Reference genome
Everything is measured against a reference
Genomics compares a sample against shared reference resources: a coordinate system, plus frequency databases that say how common each variant is. Those databases are how a variant gets called common or rare.
- 04The representation gap
A reference is a sample, not a census
Reference genomic resources were historically built from a limited set of populations. When a frequency drawn from those populations is read as if it described everyone, the number can be materially off for a population that was never in the sample.
- 05Why India
Over 1.4 billion people, and deep internal structure
India is roughly a fifth of humanity, and its diversity is highly structured rather than uniform. A single "Indian" average is already a simplification — and a global average is a further one.
The gap
A reference is asked to speak for far more diversity than built it
Nothing here is a criticism of those datasets — they are foundational. The problem starts when their numbers are reused as though they described every population.
populations the reference was built fromdiversity it is asked to represent
A frequency is only true of the group it came from
How common a DNA variant is differs between populations. Read a global average as if it described an Indian population and the number can be materially off.
The error travels downstream
Variant filtering, reference panels, imputation, association studies, and risk scores all inherit the population they were built on. Findings do not automatically transfer.
What that difference looks like on one position
Illustrative example — not real data
Made-up numbers, used only to explain how the comparison reads.
Global reference population
example
A
G
- A80%
- G20%
Indian population
example
A
G
- A60%
- G40%
Reading it: at one position in the DNA, some people carry an A and some carry a G. The bar shows how common each letter is in that group of people.
Why it matters: here the G version is twice as common in the Indian group (40%) as the global reference suggests (20%). A study, risk score, or filter tuned to the global number would treat G as rarer than it actually is in that population.
How it works
A research workflow, not a fact sheet
BharatGenome is a population-aware genomic research interface. It compares Indian and broader global genomic evidence and reports representation differences with their provenance and limits attached.
Search
a gene symbol, a variant, or an rsID
Compare
Indian population evidence against global reference evidence
Understand
what the difference means, in plain language
Check
supporting literature and where every number came from
Judge
confidence, coverage gaps, and stated limitations
01
Query
gene, variant, or rsID
02
Provider adapters
one contract per external source
03
Comparison layer
Indian vs global, on one scale
04
Provenance + confidence
source, version, and coverage attached
05
Interpretation
plain-language reading, with its limits
Why it matters
What better representation can enable
All of this is about research quality, not clinical outcomes. BharatGenome makes no claim about any individual's health.
Better-calibrated frequency baselines
When a frequency is reported against the population it was measured in, common and rare stop being guesses borrowed from somewhere else.
More relevant reference panels
Analyses that lean on reference panels — imputation, filtering, association testing — depend on those panels resembling the people being studied.
Research that transfers
Stating which population a result came from, and how well covered that population is, is what lets the next study know whether the result applies to it.
What this is
- A research and educational prototype.
- A comparison layer over genomic providers, with provenance and confidence reported alongside every result.
- An explanation of why population representation changes what a genomic finding means.
What this is not
- Not a diagnostic or clinical decision support tool.
- Not a source of treatment advice or disease prediction.
- Not a store of individual-level genetic data — BharatGenome holds none.
Current status: provider integrations are not yet connected
This build ships the product, the comparison architecture, and the provider contracts. No genomic datasets are bundled in this repository and no scientific results are invented. Until a real provider is configured, Explore reports that plainly rather than showing a number it cannot stand behind. See Methodology for exactly where real data will plug in.
Start with a gene or variant, or learn the ideas first.