What 254 Rice Varieties Reveal About Hidden Genetic Diversity
Researchers in Madhya Pradesh scored 254 rice accessions on 30 standardised traits and found that while nearly every plant looked the same at the leaf, the length and colour of the awn told them apart better than anything else.
A team at JNKVV, Jabalpur scored 250 farmer-maintained rice landraces plus four check varieties on 30 official DUS descriptors covering leaf, culm, panicle, awn, and grain traits. Basic vegetative traits like leaf collar and ligule were completely uniform across every single accession, but awn length, awn colour, and grain-colour traits showed the widest variation in the study — a practical shortlist of characteristics breeders and gene-bank curators can use to tell these varieties apart without lab equipment.
The problem
Hundreds of farmer varieties, no formal record
Original research by Vikash Sharma, Yogendra Singh, S. K. Singh, Stuti Sharma, Kanchan Bhan, and Chandrabhan Ahirwar — Department of Genetics and Plant Breeding, College of Agriculture, JNKVV, Jabalpur.
Indian farmers have kept their own rice varieties going for generations, selecting seed by hand, season after season, long before formal breeding programmes existed. These landraces carry real value — genes for drought tolerance, grain quality, and local adaptation that modern high-yielding varieties often lack. But that value is only useful to breeders and conservationists if the varieties are actually described somewhere: which ones look alike, which ones are genuinely distinct, and which traits reliably tell them apart. For a large share of India's rice landraces, particularly across Madhya Pradesh, that formal description has never been done.
The method
A standard 30-question checklist
Rather than inventing a new method, the researchers used the same descriptor system India's Protection of Plant Varieties and Farmers' Rights Authority (PPV&FRA) requires for official variety registration — the DUS test, short for Distinctness, Uniformity, and Stability. It scores plants on 30 qualitative traits, from leaf shape to grain colour, each with a fixed set of possible categories (for example, culm attitude can only be recorded as erect, semi-erect, or open). Over one growing season at JNKVV's Jabalpur research farm, five plants from each of 254 accessions — 250 indigenous landraces plus four reference check varieties — were scored on all 30 traits. To turn those raw category counts into a single diversity number per trait, the team applied the Shannon-Weaver Diversity Index (SWDI), a measure borrowed from ecology that goes to 0 when every plant falls into one category and rises as plants spread more evenly across several categories.
What stayed uniform
What barely varied at all
Two traits came back completely fixed: every one of the 254 accessions had a visible leaf collar and a visible leaf ligule, giving both an SWDI of exactly 0.00. Leaf auricle presence was almost as uniform, showing up in 98.81% of accessions. These aren't useless findings — knowing a trait is essentially invariant across regional landraces means it's not worth spending time scoring for varietal identification, and it tells conservationists these particular vegetative structures are under strong stabilising pressure regardless of which landrace they're looking at.
What set them apart
Where the real differences hide
The trait that separated these rice varieties more than any other single characteristic was, somewhat unglamorously, how long the awn — the bristle-like extension at the tip of the grain — grew. Length of the longest awn scored an SWDI of 1.55, the highest of all 30 traits measured, with accessions spread across five categories from very short to very long. Close behind came awn colour (1.43, spanning brown, reddish-brown, purple, and black among awned plants), spikelet colour at the lemma tip (1.41, white and brown predominating but with yellowish, purple, and red forms present too), and overall lemma-and-palea colour (1.40, mostly straw-coloured but including gold, brown, and purple forms). Panicle architecture told a similar story: branch attitude alone reached an SWDI of 1.28, split fairly evenly across semi-erect (40.94%), erect (30.70%), erect-to-semi-erect (17.32%), and smaller spreading and semi-erect-to-spreading classes.
Timing matters
Timing changes what you see
One of the more instructive results in the paper has nothing to do with which trait was most diverse, and everything to do with when it was measured. Flag-leaf attitude was scored twice — once early, near the start of flowering, and again later, near ripening. At the early stage, the trait showed moderate diversity (SWDI = 0.89), dominated by erect (51.96%) and semi-erect (40.94%) forms with no drooping types recorded at all. By the late-season observation, the same trait had climbed to an SWDI of 1.32, because two entirely new categories — horizontal (16.92%) and deflexed (12.59%) — had appeared as leaves aged and drooped differently across accessions. The diversity was there all along; it just wasn't visible until later in the season, a detail that matters for anyone designing a DUS scoring protocol and deciding when in the crop cycle to walk the field.
Why it matters
Turning a trait list into a breeding tool
None of this is diversity for its own sake. The PPV&FRA's DUS system exists specifically to register and legally protect distinct varieties, and a documented set of highly variable, easy-to-score qualitative traits is exactly what makes that registration process practical for landraces that have never been through it. The same trait list doubles as a starting point for breeders: picking parental lines that differ sharply in awn length, awn colour, or panicle architecture is a reasonably reliable proxy for picking genetically distant parents, without needing DNA markers to confirm it. The authors are careful to flag the limits of that shortcut, though — this was one growing season at one location, so environmental effects on trait expression couldn't be separated from genuine genetic differences, and the study measured visible morphology only, not underlying molecular diversity.
India holds an enormous, largely undocumented reservoir of farmer-maintained rice diversity, and Madhya Pradesh's landraces are no exception. Formal DUS characterisation is the legal and practical gateway to registering these varieties, conserving them deliberately rather than by accident, and using them as parents in breeding programmes aimed at climate resilience and grain quality. A concrete, ranked list of which traits actually vary — and which ones don't — turns "254 unlabelled varieties in a field" into a usable genetic resource.
What is a DUS descriptor and why does it matter for rice landraces?
DUS stands for Distinctness, Uniformity, and Stability — a standardised set of 30 qualitative traits prescribed by India's Protection of Plant Varieties and Farmers' Rights Authority for describing and legally registering crop varieties. Applying it to farmer-maintained landraces gives them a formal, comparable description they otherwise lack.
What is the Shannon-Weaver Diversity Index, and what does a score like 1.55 mean?
It is a diversity measure borrowed from ecology that scores 0.00 when every individual falls into a single category (no variation) and rises as individuals spread more evenly across multiple categories. In this study, scores ranged from 0.00 (leaf collar and ligule, fully uniform) up to 1.55 (length of the longest awn, the most variable trait measured).
Which traits were most useful for telling these rice varieties apart?
Awn length (SWDI 1.55), awn colour (1.43), spikelet colour at the lemma tip (1.41), and lemma-and-palea colour (1.40) showed the highest diversity, alongside panicle branch attitude (1.28) and panicle curvature (1.22). Leaf collar, ligule, and auricle presence were the most uniform and least useful for distinguishing accessions.
What are the limitations of this study?
The evaluation covered one growing season at a single location (JNKVV, Jabalpur), so environmental effects on trait expression could not be separated from genuine genetic differences. The study also measured visible qualitative morphology only — it did not assess molecular or DNA-marker-based diversity, which the authors recommend as a complementary next step.
Based on the peer-reviewed paper “Qualitative Morphological Characterisation and Assessment of Genetic Diversity among Indigenous Landraces of Rice (Oryza sativa L.)”. Read the full abstract, key findings, and download the PDF on the paper's own page.