Nitin Pai: The diversity of Indian genomes should cue changes in India’s approach to public health

After sequencing the whole genomes of 10,000 Indians and analysing 20,000 blood samples of individuals from over 80 communities, the research team—comprising scientists from 20 institutes across the country—has provided us with a dramatically different basis on which to think about how India manages public health.

Two preprints published by GenomeIndia researchers this year confirm that Indians are both genotypically and phenotypically hyper-diverse and the two track each other. The Indian population comprises thousands of persistently endogamous communities, loosely called jaatis.

The same process has produced thousands of distinct metabolic profiles at the clinical level. Consequently, it is not very meaningful to think of public health for Indians as a single population, or 28 state-level populations, or even 800 district-level populations. Instead of administrative boundaries, health is better mapped along community lines.

It’s striking that 95% of the sampled Indians have at least one abnormal marker. By standard clinical criteria, they would be flagged as having some metabolic problem—low HDL, high triglycerides, elevated waist circumference, high blood pressure or some combination of these.

The authors themselves pause over this number and ask whether it represents a true epidemic or an artefact of using European reference ranges on Indian bodies. Their answer: probably both.

Indians carry a heavier metabolic burden than Europeans, but the thresholds used to diagnose that burden were calibrated on a different population. So some people among the 95% may have a health risk, while it is a statistical misclassification for others.

Further, normal levels of blood sugar, cholesterol, lipids and other indicators need to be community-specific. If you get a test done today, chances are that the ‘normal’ range of that indicator comes from studies done elsewhere in the world. This is likely to be inaccurate. Even using Indian averages would be inaccurate because every community has a different norm.

The GenomeIndia project studied only 80 communities of the thousands in the country. Even this small subset shows incredibly high levels of diversity. The team identified roughly 130 million genetic variants. Of these, about 44 million had never been recorded before.

The diversity is enduring. There were 23 communities with stronger endogamy than even among Ashkenazi Jews. Genetic fingerprints of an unidentified ancient South Indian population have been found in our population. We are likely to find even more diversity if we sample more communities.

Since each community has its own normal levels of indicators and pattern of health risks, we need to establish them first. Increasingly granular genomic data should be the infrastructure on which we make public health strategies.

For instance, women with tribal ancestry are more at risk from cholesterol-related conditions than individuals with non-tribal backgrounds. A particular genetic variant linked to a metabolic disorder is present in 12.5% of one Dravidian tribal population, yet essentially absent everywhere else. Within the same district, some communities might be more prone to diabetes and others to cardiovascular risks.

Having better information that overlays disease risks onto ethnography could help achieve better and more efficient public health outcomes.

Another important finding is that while the National Family Health Survey reports diabetes prevalence at around 4.4%, the GenomeIndia study using actual blood tests rather than self-reporting finds it at 18.4%.

Hypertension is similarly undercounted. Of every 100 Indians whose blood tests show dyslipidemia, only two know they have it. Of every 100 measured as hypertensive, only 18 are aware of their condition. Such findings need deeper study but there is a serious possibility that we have thus far been unaware of the intensity of our real health problems.

What next? First, treat the GenomeIndia database as public health infrastructure. Expand it to more communities, enable longitudinal follow-ups and integrate it with existing health surveys.

Second, design community-targeted preventive screening. If a specific community is known to carry a high-frequency variant for a serious recessive disorder, offering voluntary pre-marital or newborn screening could help an entire generation fight off the effects.

Third, build an industrial base. India-specific genotyping chips, diagnostics tuned to Indian biology and drug dosing guidelines informed by Indian pharmacogenomics are achievable within five years, given existing data. Iceland built a biotech industry on a far smaller national genome project.

Fourth, today’s framework cannot promote the investment and innovation that are necessary to address the sheer diversity of tests, screenings and treatments called for and possible with the advent of inexpensive genomics and AI. We need a 1992-type liberalization of the bio-economy.

The author is co-founder and director of The Takshashila Institution, an independent centre for research and education in public policy.