Variant interpretation for Indian genetics labs, built on Indian allele data
TL;DR Allelward is software that helps lab geneticists classify the variants a sequencing test finds, using Indian population frequencies instead of mostly European ones. Clinical genetics labs and hospital genetics units pay per case or per seat. The aim is fewer "uncertain" reports and faster sign-off.
When a sequencing test finds a variant, a lab scientist has to decide whether it explains the patient's illness. A big clue is how common the variant is in healthy people of the same ancestry, and Indian labs mostly don't have that clue. A 2026 review of inherited retinal disease studies in India, covering 21,158 cases, found variant reports often lean on non-Indian population databases[1].
The result? Lots of inconclusive answers. In a South Indian retinal cohort, 41.3% of families got an inconclusive result[2], and in an Indian early-onset Parkinson's study about 45% of participants carried a variant of uncertain significance[3]. Population differences are real too - one variant common in Europeans is 98 times rarer in India[4].
What it does, in regulatory terms: decision support for qualified lab geneticists, who sign every report. It does not diagnose and does not talk to patients. Software with a medical purpose can fall under CDSCO's medical device software guidance, so classify it with a regulatory adviser before selling.
Allelward takes a lab's annotated variant file and shows each variant with Indian frequency data, published evidence and the scoring criteria labs already use. It drafts a classification and the reasons for it, and the scientist accepts or changes it. Every change is logged, which builds a shared, consented record of Indian classifications over time. That record is the real asset.
The seed pitched AI that predicts outcomes from genomic data. I reframed it because prediction from genomes is unproven for most conditions and would be a high-risk medical claim. Interpretation support for the labs already doing the tests is a narrower job with a real buyer.
Users are lab scientists and genetic counsellors. Payers are private genetics labs, hospital genetics departments and diagnostic chains running sequencing. Global tools exist, but the best-known variant search platform doesn't publish prices for its clinical product[5]. I found no public figure for what Indian labs pay. Expect per-case pricing for small labs and seat licences for larger ones.
GenomeIndia published its first national results in a leading genetics journal in April 2025[6], and studies now draw on its aggregate data of 9,768 people[4]. CDSCO finalised its medical device software guidance in July 2026, which gives you a clearer path to a licence[7].
There's no public count of Indian clinical genetics labs or their software spend. The narrow signal is the size of the uncertain-result problem in Indian cohorts[2][3] and the evidence that foreign reference data misleads[1][4]. Treat any global genomics market figure as adjacent context only.
One Bengaluru genomics firm runs its own lab with an in-house interpretation and reporting platform[8]. A Swiss variant search platform offers a free variant search used by over 500,000 professionals and a CE-IVD clinical platform[5]. Large Indian labs build their own tools. The gap is an Indian-data-first tool that independent labs can buy without hiring a bioinformatics team.
| Difficulty | Hard (56/100) |
| Starting capital | ₹1.5–2.8 Cr |
| Time to launch | 12–18 months |
The hard parts are data access and trust. Indian population data comes under controlled access, so you have to apply and accept its terms. Labs will only rely on drafts after side-by-side checks against their own signed reports, and regulatory classification may add a licence and clinical performance evidence on top.
Yes, this may need a licence. Software with a medical purpose can be a regulated device in India.
Building this outside India? The categories above usually have a local equivalent, so check with your own regulator.
Don't let AI classify a variant on its own. It's decision support, and a qualified geneticist signs every report.
Want to work out the AI layer for Allelward? Talk to Yash →
I think it's a good fit if you have clinical genomics or bioinformatics depth and lab contacts. If you're hoping AI alone will classify variants then look elsewhere. Prove first that, on 200 past cases from two labs, Allelward's drafts match signed reports and turn some uncertain results into clear ones without any reversals.
Plan for roughly ₹1.5–2.8 Cr, mostly on bioinformatics talent, data access, validation against signed reports and regulatory work.
Possibly. CDSCO's July 2026 guidance can class medical-purpose software as a device, so get a written classification view before selling.
Yes, but the edge here is Indian population data. Elsewhere you'd need local data and must meet that country's medical device software rules.
The free Allelward validation kit has the line-by-line capital breakdown, the effort scoring, the software-licensing and data-protection checklist, a 14-day validation sprint with pass/fail targets, a 90-day launch plan, 10 interview questions for lab heads, and 5 assumptions to test. Download the kit →
Thinking about building Allelward or something close to it? The make-or-break work is securing data access and proving drafts against signed reports, and both need a plan before you raise. If you need a hand, you know where to find me. Talk to Yash →
For illustration only. Names are fictional and ideas hypothetical; sourced facts are real but may be out of date. Nothing here is legal, financial or professional advice, and Yash Khatri is not personally liable for any use of it. Read the full disclaimer.