Summary
A bioRxiv preprint combines data from eight healthy Indian cohorts into a reference atlas and proposes a four-part Indian Gut Health Index. The authors report geographic patterns in the data, while noting that cohort and geography effects are strongly intertwined.
A bioRxiv preprint from researchers at BioMeBar presents a harmonised reference atlas of healthy gut microbiome data from India and proposes a population-calibrated score called the Indian Gut Health Index (IGHI). The analysis brings together eight published cohorts, covering 1,672 samples from 1,456 people across 20 locations and six geographic zones.
The atlas is intended to provide an Indian reference for a field in which microbiome databases have largely been built from Western and East Asian populations. The authors argue that applying reference ranges from other populations could misclassify Indian individuals. Their work is a preprint: its findings have not yet been peer reviewed.
How the atlas combines the cohorts
The researchers mapped data from different studies onto a shared bacterial-family taxonomy. Where raw 16S rRNA sequencing data were available, they reprocessed the sequences using a common QIIME 2, DADA2 and SILVA 138.1 pipeline. For cohorts available only as taxonomic tables, they retained the original classifications and mapped them to the shared taxonomy through a series of matching and fallback steps.
The resulting matrix contained 380 bacterial families. Seventeen were present in at least 80% of samples and were designated the pan-Indian core. Lachnospiraceae appeared in all samples, while Prevotellaceae appeared in 99.5%.
The authors also identified two broad community types: one dominated by Prevotellaceae, accounting for 69% of samples, and another dominated by Oscillospiraceae, accounting for 31%. They describe this pattern as different from the three-enterotype model commonly associated with Western populations. These categories summarise patterns in the combined dataset; they are not, by themselves, diagnoses or measures of an individual's health.
What the geographic patterns mean
The analysis found significant differences in microbiome composition across both locations and cohorts. However, the two influences were difficult to separate: variance partitioning attributed 30.4% of variation to effects shared between geography and cohort, compared with 4.1% to geography alone and 4.7% to cohort alone. After accounting for cohort in restricted permutation tests, the city-level geographic signal remained significant, while the zone-level signal did not.
A machine-learning classifier assigned samples to geographic zones with 68.4% balanced accuracy, compared with a reported null value of 16.6%. The Northeast group was most distinguishable, with an F1 score of 0.92. These results indicate that some location-related patterns were detectable in this dataset; they do not separate all regional differences from the way the contributing studies were designed and collected.
The IGHI combines four components: microbial diversity, taxon balance, membership in the core, and how typical a community is relative to geography-stratified healthy Indian norms. Reported zone-average scores ranged from 67.6 to 72.9, a 5.3-point spread. The authors also report that rankings remained robust under five of six alternative weighting schemes, with Spearman correlations of at least 0.88.
The atlas and score offer a framework for comparing microbiome profiles against Indian healthy-cohort data. The preprint derives the score from healthy reference norms; it does not report clinical diagnostic validation. Differences in original study methods, together with the strong overlap between cohort and geography effects, are important when interpreting the atlas as a population reference.