Why 23andMe Says "Broadly South Asian" — And How to Get Real Indian Ancestry Results

You spent ₹10,000-20,000 on a 23andMe test, waited 6-8 weeks for results, and opened your ancestry report expecting to learn something fascinating about your heritage. Instead, you got: "97.2% South Asian. 2.8% Broadly South Asian." Maybe, if you were lucky, the 2025 update gave you "Northern Indian and Pakistani" or "Bengali and Northeast Indian." But for most Indian users, the experience is the same: vague categories that tell you nothing you did not already know.

You are not alone. This is the single most common complaint from Indian users of 23andMe, AncestryDNA, and MyHeritage. And the reason it happens is not a bug or a limitation of DNA testing technology. It is a reference panel problem, a fundamental design choice that makes global tests almost useless for the 1.4 billion people who live in the most genetically diverse country on Earth.

Understanding the "Broadly South Asian" Problem

To understand why 23andMe fails Indians, you need to understand how ancestry estimation works at a technical level. Every DNA ancestry test follows the same basic process: compare your DNA against a reference panel of known populations, and calculate which populations your DNA most closely resembles.

How Reference Panels Work

A reference panel is a curated database of genomes from people whose ancestry is well-documented, typically going back at least three generations. When 23andMe says you are "42% British and Irish," they are comparing your DNA to genomes from people who have four grandparents all born in Britain or Ireland. The more reference samples you have from a specific population, the more confidently the algorithm can assign ancestry to that group.

Here is where the problem for Indians begins. 23andMe's reference panel, as of early 2025, contained approximately:

• Europeans: ~10,000+ reference samples across 40+ sub-populations (British, Irish, French, German, Italian, Greek, Scandinavian, etc.)
• East Asians: ~3,000+ reference samples across 20+ sub-populations (Chinese, Japanese, Korean, Filipino, Vietnamese, etc.)
• Africans: ~2,000+ reference samples across 30+ sub-populations
• South Asians: ~150-300 reference samples across 5-8 sub-populations

The disparity is staggering. Europe, with a population of 450 million and relatively low genetic diversity (due to extensive historical gene flow between European populations), gets 10,000+ reference samples and 40+ sub-populations. India, with a population of 1.4 billion and the most genetically structured population on Earth (with over 4,500 endogamous communities), gets fewer than 300 reference samples and 5-8 sub-populations.

Why India Is Uniquely Hard for Global Tests

India is not just another underrepresented region. It presents a unique challenge for ancestry algorithms because of three factors that make it fundamentally different from Europe or East Asia.

Extreme endogamy: Indian communities have practiced endogamy (marriage within the community) for approximately 1,500-3,000 years. This has created thousands of genetically distinct population clusters. A Nair from Kerala is as genetically distinct from a Jat from Haryana as a Spaniard is from a Finn. But where 23andMe can distinguish Spanish from Finnish with ease (because it has hundreds of reference samples from each), it cannot distinguish Nair from Jat because it lacks sufficient reference samples from either community.

The ANI-ASI cline: Indian genetic variation follows a primary north-south gradient called the ANI-ASI cline. Northern populations have more Ancestral North Indian (ANI) ancestry (related to Central Asian and Iranian populations), while southern populations have more Ancestral South Indian (ASI) ancestry (indigenous to the subcontinent). This cline is continuous, not categorical. Without a dense reference panel that samples communities along the entire gradient, an algorithm cannot determine where on the cline a person falls, so it defaults to "South Asian."

Multi-dimensional variation: Unlike European genetic variation, which is primarily captured by a single north-south/east-west gradient, Indian genetic variation is multi-dimensional. Beyond the ANI-ASI axis, there is an East Asian admixture gradient (highest in Northeast and Bengal), an Austro-Asiatic substrate (in central and eastern tribal populations), and extensive community-specific drift. Capturing this complexity requires a reference panel that is not just large but specifically designed for Indian genetic architecture.

23andMe's 2025 South Asian Update: Better, But Still Inadequate

In 2025, 23andMe rolled out a significant update to their South Asian ancestry reporting. The previous single "South Asian" category was replaced with seven sub-regions:

1. Northern Indian and Pakistani
2. Southern Indian
3. Bengali and Northeast Indian
4. Gujarati
5. Sri Lankan
6. Central Asian and Northern Indian
7. Broadly South Asian

This was unquestionably an improvement. Users who previously saw only "100% South Asian" could now see, for example, "62% Northern Indian and Pakistani, 23% Gujarati, 15% Broadly South Asian." For some users, particularly Gujaratis (who were well-represented in 23andMe's customer base), the update provided meaningful new information.

Why 7 Sub-Regions Are Not Enough

But 7 sub-regions for 1.4 billion people is still woefully inadequate. Consider what each category actually covers:

23andMe Category	Populations Lumped Together	Approximate Population
Northern Indian and Pakistani	Punjabis, Rajputs, Jats, Khatris, Brahmins (UP/Bihar/MP), Sindhis, Pashtuns, Baloch, Kashmiris, etc.	~600 million
Southern Indian	Tamils, Telugus, Kannadigas, Malayalis, all castes and tribes within each	~350 million
Bengali and Northeast Indian	Bengalis (all castes), Assamese, Nagas, Mizos, Manipuris, Khasis, etc.	~250 million
Gujarati	Gujarati Brahmins, Patels, Jains, Lohanas, Kolis, Bhils, Adivasis, etc.	~70 million
Broadly South Asian	Everyone the algorithm cannot confidently assign to the above categories	Varies

Telling a Rajput from Udaipur that they are "Northern Indian and Pakistani" provides almost no information. They already knew they were from North India. They wanted to know they are specifically Rajasthani with strong Marwari community affinity, elevated steppe ancestry characteristic of Rajput clans, and R1a-Z93 haplogroup connections to Bronze Age Central Asian migrations. That level of detail requires a reference panel with thousands of Indian communities, not 7 broad categories.

The persistent "Broadly South Asian" category is particularly frustrating. For many Indian users, 15-40% of their ancestry remains classified as "Broadly South Asian" even after the 2025 update. This is the algorithm's way of saying "I know you are from South Asia, but I do not have enough reference data to be more specific." It is the genetic equivalent of a shrug.

What a Real Indian DNA Test Shows

To understand the gap between global and Indian-specific tests, let us look at what a typical Indian user receives from each service.

Typical 23andMe Result for an Indian User

• 62.3% Northern Indian and Pakistani
• 18.7% Broadly South Asian
• 12.1% Southern Indian
• 4.2% Central Asian and Northern Indian
• 2.7% Unassigned

Total meaningful Indian ancestry detail: effectively zero beyond "you are mostly North Indian with some South Indian." The person already knew this from their family history.

Typical Helixline Result for the Same Person

• Ancestry Composition: 34% Rajasthan (Marwari cluster), 22% Gujarat (Patel-adjacent), 18% Madhya Pradesh (Malwa region), 14% Uttar Pradesh (Awadhi cluster), 8% Maharashtra (Vidarbha), 4% other regions
• Community Match: Highest affinity to Rajasthani Rajput communities (93% confidence), secondary affinity to Gujarati Vania communities (78% confidence)
• ANI/ASI Breakdown: 68% ANI, 30% ASI, 2% East Asian-related
• Paternal Haplogroup: R1a-Z93 > L657.1, associated with Bronze Age steppe ancestry common in Rajput and upper-caste North Indian populations
• Maternal Haplogroup: U2e1, a lineage common in western Indian populations with deep roots in the subcontinent
• Neanderthal DNA: 1.67% (higher than 72% of Indian users tested)
• Migration Path: Steppe > Iranian Plateau > Indus Valley > Rajasthan > Gujarat (with estimated timeline)

The difference is not incremental. It is transformative. One result tells you nothing new. The other tells you a detailed story about who your ancestors were, where they lived, and how they moved across the subcontinent over thousands of years.

The Technical Explanation: Why Resolution Matters

For readers who want to understand the technical underpinnings, here is why reference panel size directly determines ancestry resolution.

Principal Component Analysis and Clustering

Ancestry estimation algorithms use techniques like principal component analysis (PCA) and model-based clustering (ADMIXTURE) to assign your DNA to populations. In PCA, each genome is plotted in a high-dimensional space defined by genetic variation. Populations that are genetically similar cluster together. The algorithm assigns your ancestry based on which reference clusters your DNA falls closest to.

With only 7 South Asian reference clusters, the algorithm can distinguish a Bengali from a Gujarati (these populations are genetically quite distant), but it cannot distinguish a Rajasthani Rajput from a Uttar Pradesh Thakur (these populations are genetically close and require dense reference sampling to separate). The result is that closely related but distinct populations get lumped into the same broad category.

With 4,500+ reference clusters (as in Helixline's panel), the algorithm has enough resolution to separate populations that differ by even small amounts of genetic drift. This is why Helixline can identify your ancestry at the community level while 23andMe can only manage broad regional categories.

The F-statistic Problem

Population geneticists use a metric called F_ST (fixation index) to measure genetic differentiation between populations. For context, F_ST between British and Italian populations is approximately 0.004-0.006. F_ST between a Rajasthani Brahmin and a Tamil Paraiyar is approximately 0.04-0.06, roughly 10 times higher. Indian populations are dramatically more differentiated from each other than European populations are.

This means that a reference panel that is sufficient for European ancestry estimation (able to distinguish populations with F_ST of 0.004) is completely inadequate for Indian ancestry estimation (where the relevant F_ST values range from 0.001 for closely related communities to 0.06 for the most divergent Indian groups). To achieve the same resolution in India that 23andMe achieves in Europe, you need proportionally far more reference samples, not fewer.

How to Get Better Results from Your Existing 23andMe Data

If you have already taken a 23andMe test and are disappointed with your South Asian results, there are several options to extract more value from your existing data.

Option 1: Upload to Helixline (Coming Soon)

Helixline is developing a raw data upload feature that will allow you to import your 23andMe (or AncestryDNA/MyHeritage) raw data file and reanalyze it against Helixline's Indian-specific reference panel. This will provide state-level ancestry, community matching, and ANI/ASI component analysis using your existing genotype data. Check helixline.in for the latest availability.

Important caveat: raw data uploads will not provide the full resolution of a native Helixline test. This is because 23andMe's genotyping chip is designed for European-relevant genetic variants and may not cover all the South Asian-specific variants that Helixline's chip captures. You will get significantly better results than 23andMe's own analysis, but for the best possible Indian ancestry detail, a fresh Helixline test is recommended.

Option 2: Use Third-Party Analysis Tools

Several free and paid tools allow you to upload raw DNA data for additional analysis. GEDmatch provides admixture calculators (such as Harappa World and MDLP) that include South Asian-specific reference populations. These tools can give you a rough breakdown into components like "Baloch," "South Indian," "NE Euro," and "SE Asian," which are more informative than 23andMe's broad categories. However, these tools require significant genetic literacy to interpret correctly and can produce confusing or misleading results for non-experts.

Option 3: Order a Helixline Test

The most straightforward solution is to order a purpose-built Indian DNA test. Helixline's Ancestry Essential kit at ₹6,999 provides the level of detail that Indian users expect and that global tests cannot deliver. If you have already spent ₹16,000-20,000 on a 23andMe test from India, spending ₹6,999 on a test that actually tells you something meaningful about your Indian heritage is a worthwhile investment.

Side-by-Side Comparison: What You Get for Your Money

Feature	23andMe (from India)	Helixline Ancestry Essential
Price (effective, from India)	₹16,000-20,000	₹6,999
South Asian sub-regions	7 broad categories	28 states + 4,500+ communities
Community matching	Not available	4,500+ Indian communities
ANI/ASI breakdown	Not available	Included
Indian-specific haplogroups	Basic haplogroup only	Detailed subclades with Indian context
Migration pathway	Generic "South Asian"	Detailed subcontinental migration
Ships to India	No (requires forwarding)	Yes, free two-way shipping
DNA Relatives matching	Yes (large global database)	Growing Indian database
Reference panel for Indians	~150-300 samples	4,500+ community samples

The one area where 23andMe retains a clear advantage is its DNA Relatives feature. With over 14 million customers, 23andMe has the world's largest genetic database for relative matching. If finding distant cousins is your primary goal, particularly in the diaspora, 23andMe's relative database is unmatched. Many Indian users choose to take both tests: 23andMe for the relative matching and Helixline for actual ancestry detail.

The NRI Perspective: Special Considerations for Indians Abroad

For Non-Resident Indians (NRIs) in the US, UK, Canada, Australia, or the Gulf, the situation has some nuances worth considering.

23andMe Is Easier to Access Abroad

If you live in a country where 23andMe ships directly, the logistical advantage is real. You order online, the kit arrives in days, you spit in the tube, drop it in the mail, and results arrive in 3-4 weeks. There is no international shipping hassle. For this convenience, some NRIs are willing to accept the limited South Asian resolution.

But Ancestry Curiosity Runs Deeper for NRIs

Paradoxically, NRIs often have a deeper hunger for detailed ancestry information than Indians living in India. When you live in your ancestral homeland, surrounded by family and community, your identity is reinforced daily. When you live in Houston or London or Dubai, your Indian identity becomes more abstract, and the desire to understand it at a molecular level becomes more intense. Being told "you are South Asian" by 23andMe can feel almost insulting when you were hoping to discover the specific threads that connect you to your ancestral village.

Helixline ships internationally, making it accessible to the global Indian diaspora. For NRIs who have already tried 23andMe and been disappointed, Helixline offers the detail they were looking for in the first place.

Frequently Asked Questions

Why does 23andMe say I am 100% South Asian with no breakdown?

23andMe's ancestry algorithm relies on a reference panel that historically contained fewer than 200-300 Indian samples for 1.4 billion people. Without enough reference samples from different Indian regions and communities, the algorithm cannot distinguish between genetically distinct populations. Even with the 2025 update adding 7 South Asian sub-regions, many users still see a large "Broadly South Asian" percentage because the reference panel remains too sparse to confidently assign ancestry to specific categories. Indian-specific tests like Helixline use panels with 4,500+ Indian community samples, providing the resolution needed for meaningful results.

Did 23andMe improve South Asian results in their 2025 update?

Yes, the 2025 update replaced the single "South Asian" category with 7 sub-regions: Northern Indian and Pakistani, Southern Indian, Bengali and Northeast Indian, Gujarati, Sri Lankan, Central Asian and Northern Indian, and Broadly South Asian. This was a notable improvement, particularly for Gujarati users who had strong representation in 23andMe's customer base. However, 7 sub-regions for 1.4 billion people remain extremely broad. "Northern Indian and Pakistani" alone covers over 600 million people across dozens of states and thousands of communities. For most Indian users, the results are still too vague to be meaningful compared to tests with Indian-specific reference panels.

Can I upload my 23andMe data to Helixline for better Indian results?

Helixline is building a raw data upload feature that will allow users to import 23andMe, AncestryDNA, or MyHeritage raw data files for reanalysis against Helixline's Indian-specific reference panel. This will provide state-level ancestry, community matching, and ANI/ASI analysis using your existing data. For the most comprehensive results, a fresh Helixline test is recommended, as Helixline's genotyping array is optimized for South Asian genetic variants that 23andMe's chip may not cover. Check helixline.in for the latest availability of the upload feature.

What does a real Indian DNA test show that 23andMe does not?

An Indian-specific test like Helixline provides ancestry composition at the state level (for example, 34% Rajasthan, 22% Gujarat, 18% Madhya Pradesh), community matching across 4,500+ Indian endogamous groups with confidence scores, ANI and ASI component percentages, detailed haplogroup analysis with Indian-specific subclade information and historical context, migration pathway mapping showing ancestral movements across the subcontinent, and Neanderthal/Denisovan ancestry calibrated for South Asian populations. A typical 23andMe result for the same person would say "62% Northern Indian and Pakistani, 19% Broadly South Asian."

Is 23andMe worth it for Indians living abroad?

23andMe has some value for NRIs: its DNA Relatives feature with 14+ million customers is unmatched for finding distant cousins in the diaspora, and its health reports are well-validated with FDA approval for select conditions. However, for ancestry purposes, the results remain disappointingly vague for most Indians. Many NRIs take both tests: 23andMe for relative matching and health features, and Helixline for actual Indian ancestry detail. If you can only afford one test and your primary goal is understanding your Indian heritage, Helixline provides dramatically more value at a lower price point (₹6,999 vs. ₹16,000+ for 23andMe from abroad).

Conclusion

The "Broadly South Asian" problem is not a flaw in DNA testing technology. It is a consequence of building tests for a European and American market and treating 1.4 billion Indians as an afterthought. With fewer than 300 reference samples representing all of South Asia, 23andMe simply does not have the data to tell Indian users anything specific about their heritage.

The 2025 update improved matters from "useless" to "slightly informative," but 7 sub-regions for the world's most genetically diverse country remains inadequate by any standard. If you are Indian and you want a DNA test that treats your ancestry with the same level of detail that European users have enjoyed for a decade, you need a test built specifically for Indian genetics.

Whether you are in India or part of the global diaspora, the era of settling for "Broadly South Asian" is over. Tests like Helixline now provide the state-level, community-level, and component-level detail that Indians have always deserved but never had access to before. Your DNA has a specific and fascinating story to tell. You just need the right test to read it.

For more on how DNA testing works for Indians, read our 2026 DNA test price comparison and our guide to understanding your Indian DNA results.