Illustrative DNA vs Helixline for South Asians
If you are South Asian, live in the US, UK or Canada, and already have a 23andMe or AncestryDNA raw data file, you have probably come across Illustrative DNA. Its charts are all over TikTok and ancestry forums: percentages of "Indus Periphery", "Steppe" and "AASI", maps of Bronze Age migrations, and lists of ancient samples you supposedly resemble. It looks scientific, and much of it is built on real scientific tools. So the obvious question is: is Illustrative DNA accurate, and is it the right upload for you?
This guide explains what Illustrative DNA does, what its ancient components mean in plain English, and why percentages move between models. We run a competing service, so we have tried hard to be fair: Illustrative DNA is good at what it sets out to do, which is simply different from what most diaspora families are asking.
The short version: Illustrative DNA models your genome as a mix of ancient and modern reference populations. It is well suited to deep-time questions (how much Steppe-related or AASI-related ancestry do I carry?) and to hobbyists who enjoy running their own models. It does not give a state-by-state South Asian breakdown, community comparisons or health reports. Helixline Upload Ancestry ($25) focuses on exactly those South Asian questions, from the same raw file.
What Illustrative DNA actually does
Illustrative DNA is an online service run by IllustrativeDNA OĆ that accepts raw data files from MyHeritage, 23andMe, AncestryDNA, FamilyTreeDNA, Living DNA and TellMeGen. According to its website, it currently offers three things:
- DeepAncestry. Its flagship report. Your raw data is projected onto a PCA-based coordinate space built from thousands of ancient and modern individuals. Those coordinates (a close cousin of the "G25 coordinates" popular with hobbyists) are used to find your closest ancient and modern populations, rank best-fitting two-way and three-way mixtures, and draw PCA plots, organised by era and including an "Ancient South Asia" group.
- AdmixLab. A subscription, do-it-yourself environment for qpAdm and Fst, the same statistical tools used in academic ancient-DNA papers. Its own FAQ is refreshingly candid: there are no pre-generated results, "a basic understanding of population genetics and ancient-DNA literature is still required", and qpAdm is "a relatively hard tool to use efficiently".
- HaploMapper. A free haplogroup tool that compares your haplogroup against published ancient samples.
Price: Illustrative DNA did not list prices on the public pages we checked, and forum figures go out of date. Check their site for current pricing.
Is Illustrative DNA accurate?
The fair answer is: the methods are real, but "accurate" depends on the question you ask.
PCA coordinates, genetic distances and qpAdm are standard tools in population genetics. When Illustrative DNA says your genome sits close to a cluster of samples, that is a genuine measurement of similarity.
What an admixture model does not give you is a literal list of who your ancestors were. A model such as "60% Indus Periphery, 25% Steppe, 15% AASI" means that a mixture of those reference groups, in those proportions, reproduces your genetic position well. A different set of reference groups can reproduce it almost as well with different numbers. The service is transparent about this: it shows a "genetic fit" score and lists the top-ranked combinations rather than a single answer.
So Illustrative DNA is a reasonable way to explore deep ancestry, provided you read its percentages as best-fitting models, not as a family tree.
Steppe, Iranian farmer and AASI in plain English
Most South Asian results on ancient-population services are built from three broad ancestral streams identified by the large ancient-DNA study of the region (Narasimhan et al., Science, 2019). Here is what each means, without the jargon.
Iranian farmer related ancestry (and "Indus Periphery")
This is ancestry related to early farmers and herders of the Iranian plateau. Ancient individuals from sites connected to the Indus Valley Civilisation carried a mix of this ancestry and AASI. Researchers call that mix the "Indus Periphery Cline", which is why you often see "Indus Periphery" as a component on hobbyist reports. It is a label for a group of ancient samples, not proof that your family lived in Harappa.
Steppe (Steppe_MLBA)
This stands for Middle to Late Bronze Age Steppe ancestry: people from the Eurasian grasslands whose ancestry spread into South Asia from roughly 4,000 years ago. It is found across the subcontinent, generally at higher levels in the north and north-west. For more, see our guide to Steppe ancestry in India.
AASI (Ancient Ancestral South Indian)
AASI is the deep, local hunter-gatherer ancestry of South Asia, most closely related (distantly) to present-day Andamanese groups. No ancient individual with pure AASI ancestry has yet been sequenced, so it is modelled indirectly. Every mainland South Asian population carries some of it. Our explainer on AASI genetics goes deeper.
These streams mixed over thousands of years into the two clusters first described in Reich et al., Nature, 2009: Ancestral North Indians (ANI) and Ancestral South Indians (ASI). Almost every South Asian today is a blend of ANI and ASI, which is why reports on both services talk about ratios rather than single labels.
Why your percentages change between models
Many people run their file through Illustrative DNA, then a GEDmatch calculator, then a forum member's qpAdm spreadsheet, and get three different answers. That is expected. The main reasons:
- Different source populations. Model yourself with "Indus Periphery + Steppe + AASI" and you get one set of numbers. Swap in a different ancient Iranian sample, or split AASI into a modern proxy group, and the proportions shift, sometimes by 10 percentage points or more.
- Proximal versus distal models. A distal model uses very old, very broad sources. A proximal model uses more recent, more specific ones. Both can be valid at the same time; they simply answer different time depths.
- Limited overlap with consumer chips. Illustrative DNA's own AdmixLab FAQ says to expect roughly 100,000 to 400,000 SNPs shared between a commercial raw file and its reference datasets, depending on the company and chip version. Fewer shared markers means wider uncertainty.
- Statistical noise. Small components (say 2-3%) are often within the margin of error, especially for exotic-sounding labels.
Treat any single percentage as an estimate with a range. Patterns that stay stable across several reasonable models are the ones worth trusting.
What Illustrative DNA does not do for South Asian families
These are not criticisms of quality. They are outside its scope, and they are the questions diaspora customers ask us most often.
- No state-level South Asian breakdown. Its modern matches are reported as genetic distances to reference populations, not as a structured "Punjab / Gujarat / Kerala" style breakdown designed around the subcontinent.
- No community comparisons shortlist. It does not rank your similarity to a curated set of South Asian community reference groups.
- No health, wellness or pharmacogenomics reports. It is an ancestry service only.
- A steep jargon curve. Terms like "genetic fit", "distal", "Fst" and sample codes are normal in population genetics, but they leave many first-time users unsure what their result means for their own family story.
Illustrative DNA vs Helixline: side by side
| Feature | Illustrative DNA | Helixline Upload |
|---|---|---|
| Main focus | Ancient and modern population modelling worldwide | South Asian ancestry in readable detail |
| Ancient-component modelling | ā Detailed, by era, plus DIY qpAdm (AdmixLab) | ā Ancient DNA similarity (Indus Valley, Steppe, AASI references), for historical context |
| State-level South Asian breakdown | - | ā |
| Community comparisons | - | ā 140+ community reference groups with a closest-match shortlist |
| ANI / ASI composition | Via ancient models | ā Reported directly |
| Chromosome painting | - | ā Across 22 autosomes |
| Haplogroups | ā HaploMapper (free) | ā Y-DNA and mtDNA, if your file includes those markers |
| Wellness, carrier screening, pharmacogenomics | - | ā Upload Complete ($50), informational only |
| Accepted files | MyHeritage, 23andMe, AncestryDNA, FamilyTreeDNA, Living DNA, TellMeGen | 23andMe (v3, v4, v5), AncestryDNA, MyHeritage, FamilyTreeDNA, LivingDNA |
| Price | Check their site for current pricing | $25 Ancestry Ā· $50 Complete (one-time) |
| Data handling | See their privacy policy | Encrypted, stored in India under the DPDP Act 2023; raw file auto-deleted within 30 days; never sold |
Already have your raw file? Ask it a South Asian question
Upload the same 23andMe or AncestryDNA file to Helixline for a state-level breakdown, community comparisons and ANI/ASI composition. From $25, one-time, results within 1-2 days.
Upload Your Raw DataWho should use which?
Illustrative DNA is a good fit if you
- Enjoy ancient history and want to explore how Bronze Age and Iron Age populations relate to your genome.
- Are comfortable with population-genetics jargon and want to build your own qpAdm models.
Helixline is a good fit if you
- Want to know which regions of South Asia your DNA most resembles, in a report you can share with parents and grandparents.
- Are curious about similarity to community reference groups, ANI/ASI composition and your haplogroups, explained in plain English.
- Also want wellness traits, carrier screening and pharmacogenomics insights (Upload Complete, $50). These are informational and not a diagnosis.
Plenty of hobbyists use both. They answer different questions, and your raw file is not "used up" by uploading it once.
How to upload the same file to Helixline
- Download your raw data from 23andMe, AncestryDNA, MyHeritage, FamilyTreeDNA or LivingDNA. Keep it unmodified (the original .txt or .zip). Our guide to what raw DNA data is explains the format.
- Choose your plan and currency on helixline.in/upload and pay: Upload Ancestry is $25 / Ā£19 / CA$34 / ā¹1,999 and Upload Complete is $50 / Ā£39 / CA$68 / ā¹3,999 (other currencies shown at checkout). Pay with Razorpay in 8 currencies or PayPal in USD.
- Check your email for your dashboard link.
- Sign in with the same email and upload your file. It takes about 2 minutes.
- Get your results, typically within 1-2 days.
You get a full refund if you cancel before analysis starts, or if your file cannot be processed (see the upload terms). Your data is encrypted and stored in India under the DPDP Act 2023, the uploaded raw file is automatically deleted within 30 days (you can delete it sooner at any time), and your genetic data is never sold.
Frequently Asked Questions
What does "Indus Periphery / Steppe / AASI" mean?
They are labels for ancient ancestry sources used in South Asian population models. Indus Periphery refers to ancient individuals connected to the Indus Valley Civilisation, who carried a mix of Iranian farmer related ancestry and AASI. Steppe (often Steppe_MLBA) is ancestry from Bronze Age pastoralists of the Eurasian grasslands that reached South Asia from around 4,000 years ago. AASI, Ancient Ancestral South Indian, is the deep local hunter-gatherer ancestry of the subcontinent. Your percentages describe how well a mix of these reference groups fits your DNA, not a literal list of ancestors.
Why do my percentages change between models?
Because each model uses different reference populations and time depths. Swapping one ancient Iranian sample for another, or using a modern proxy instead of AASI, can move the numbers by several percentage points while still fitting your data well. Consumer chips also share only part of their markers with ancient reference datasets, which widens the uncertainty. Patterns that stay stable across several reasonable models are more meaningful than any single number.
Is ancient-population modelling reliable for one person?
It is reliable as a broad picture and less reliable in the fine detail. Tools like PCA and qpAdm are used in published research, usually on groups of samples. For a single consumer file, expect the big picture (for example, more or less Steppe-related ancestry than average) to be sound, while small components of a few percent may be noise. Read the results as estimates with ranges.
Can I use the same raw file for both?
Yes. Your raw data file is simply a text file of your genotypes, and uploading it to one service does not stop you uploading it to another. Download the unmodified .txt or .zip from 23andMe, AncestryDNA, MyHeritage, FamilyTreeDNA or LivingDNA and use it with each service. Always read each service's privacy policy before uploading.
Which is better for finding my Indian state or community?
For that specific question, Helixline is the more direct fit, because it is built to report a state-level ancestry breakdown and your similarity to 140+ South Asian community reference groups, with a closest-match shortlist. Illustrative DNA focuses on ancient and worldwide population models rather than a state-by-state South Asian breakdown. Note that no DNA test can prove membership of a caste or community; results show genetic similarity to reference groups.
Curious what your file says about your South Asian roots? Upload your raw data to Helixline from $25. If you want more background first, read ANI vs ASI: the two ancestries in every Indian's DNA, our guide to DNA upload sites for South Asian ancestry, or how to fix a "Broadly South Asian" result.