Madagascar in the Global Flora League Tables

← Back to Knowledge Hub

Two rankings. One story.

When biologists and commercial teams try to rank biodiversity hotspots for bioprospecting value, two metrics are commonly cited: the raw number of endemic species in a territory, and the density of those endemics per unit area. Both tell you something useful. Neither, on its own, tells you enough.

Madagascar sits 5th in the world by raw endemic plant count — behind Indonesia, Brazil, Australia, and China. By endemic density (endemics per 1,000 km²), it ranks 12th, trailing a string of small island nations that concentrate relatively few species across tiny land areas. If you stopped the analysis there, you might conclude that Madagascar is a significant but not exceptional biodiversity target.

The number that changes the analysis is the endemism ratio: the share of a country's total flora that is endemic to that country alone. Madagascar's endemism ratio is 83% — the highest of any large flora on Earth. Not 83% of a small collection of species. 83% of a flora of nearly 14,000 vascular plant species. That ratio is what transforms a ranking position into a commercial proposition: it means that for the chemistry encoded in these plants, there is no alternative geographic source.

The raw endemic count league table

The table below shows the top ten countries by confirmed endemic vascular plant count, drawn from the Kew World Checklist of Vascular Plants (January 2026 release, 362,739 accepted species) cross-referenced with NCBI GenBank assembly data. Madagascar appears in fifth position. The figures for endemism ratio — the percentage of each country's total flora that is endemic — reveal the structural difference.

Rank	Country	Endemic species	Endemism ratio
1	Indonesia	20,621	65%
2	Brazil	15,326	42%
3	Australia	13,111	60%
4	China	12,305	36%
5	Madagascar	9,829	83%
6	Chile	9,159	31%
7	Peru	7,420	36%
8	Mexico	7,310	29%
9	Ecuador	5,739	31%
10	Philippines	4,931	52%

Source: Kew World Checklist of Vascular Plants, Jan 2026 release. Native, non-extinct, non-introduced species only.

The ratio column is the operative variable. Indonesia has more than twice Madagascar's endemic count — but 35% of its plants are also found elsewhere, meaning there are substitute geographic sources for a substantial portion of its genetic diversity. Madagascar's 83% ratio means that for the overwhelming majority of its flora, there is no substitute. The chemistry is here and nowhere else.

"In the countries with the highest raw endemic counts, most endemics are also found somewhere else. Madagascar's endemism ratio means the opposite is almost universally true."

The endemic density league table

Endemic density — endemics per 1,000 km² of land area — is a measure of geographic concentration. It answers a different question: how tightly packed is this endemic diversity into a given area? The table below shows the top countries by this metric.

Rank	Country	Endemics / 1,000 km²	Total endemics	Land area (km²)
1	Seychelles	169.9	78	459
2	Mauritius	128.4	262	2,040
3	Comoros	94.9	212	2,235
4	Samoa	76.4	217	2,842
5	Jamaica	74.7	821	10,991
6	Fiji	43.7	799	18,272
7	Tonga	38.8	29	747
8	Costa Rica	36.3	1,857	51,100
9	Singapore	32.0	23	719
10	Cuba	25.7	2,827	109,884
11	Ecuador	20.2	5,739	283,561
12	Madagascar	16.7	9,829	587,041

Source: WCVP Jan 2026 endemic counts cross-referenced with World Bank 2022 land area data.

Why island nations dominate density — and why scale inverts the picture

The density table is dominated by small island nations and archipelagos. This is a mathematical consequence of how endemic density is calculated: a small island with 100 endemic species across 500 km² will score 200 per 1,000 km², while a large island with 10,000 endemic species across 600,000 km² will score only 16.7 per 1,000 km², even though it holds a hundred times more endemic species and a hundred times more of the associated chemical diversity.

The Seychelles ranks first with 169.9 endemics per 1,000 km². Its total endemic flora is 78 species. Madagascar ranks 12th with 16.7 per 1,000 km². Its total endemic flora is 9,829 species — 126 times larger. For any organisation planning to access endemic plant diversity at scale — building a genomics programme, sourcing metabolomics data, training AI models on novel biological chemistry — the absolute size of the available library matters as much as its density.

Density is a useful proxy for bioprospecting efficiency in a purely logistical sense: a high-density flora means endemic species are geographically concentrated. But the total number of species determines the ceiling of the scientific and commercial opportunity. Madagascar's 12th-place density ranking and its 5th-place raw count combine to produce the world's largest accessible endemic plant library within a single legal jurisdiction.

The genomic gap in global context

The rankings above measure documented taxonomic diversity. The genomic gap data adds a second dimension: of the documented endemic species, how many have been characterised at the genetic level? The answer, globally, is almost none — and Madagascar is representative of the global pattern.

Across all 362,739 accepted vascular plant species in the WCVP database, only 2,758 (0.76%) have any genome assembly record in NCBI GenBank. The highest-coverage major flora is China's, with 906 species sequenced — representing 2.64% of its total plant diversity. Most of the world's plant species have never been genetically characterised at any level.

Madagascar's coverage is 85 species with any NCBI assembly data — 0.72% of its total flora, consistent with the global average. Of its 9,829 confirmed endemics, 3 have chromosome-scale reference genome assemblies. All other standard genomic databases — NCBI GenBank, Ensembl Plants, GBIF — contain sequence data for fewer than 500 Malagasy endemic species in total, the overwhelming majority limited to barcode-grade sequences rather than functional genomic data.

The 6,080 figure — Madagascar endemic species with no genetic data recorded in NCBI — is the commercially operative number. It represents species that are, in data terms, entirely dark: no genome sequence, no transcriptome, no functional annotation, no compound-gene linkage. Each one may carry biosynthetic gene clusters encoding compound families that have never been observed in any database. The 165 million years of evolutionary isolation that produced Madagascar's extraordinary taxonomy has been producing equally extraordinary chemistry — and 6,080 of the species carrying that chemistry are completely uncharacterised.

"6,080 endemic species. No genetic data in any public database. Not one. That is the scale of the uncharacterised library."

What the tier breakdown shows

The genomic coverage of Madagascar's endemic flora is not binary — there are degrees of characterisation, from chromosome-scale reference genomes at the highest quality level down to short barcode sequences. The IsoGentiX database classifies every species against NCBI data using a four-tier system:

• Tier 3 — Chromosome-scale genome: 3 endemic species. These are the species that have received the most intensive genomic investment globally — full reference genomes assembled to Earth BioGenome Project standard or equivalent.
• Tier 2 — Draft genome: 18 endemic species. Partial genome assemblies — useful for gene discovery but not complete enough for structural genomics work.
• Tier 1 — Barcode / plastid sequences only: 3,363 endemic species. Short sequences suitable for species identification but carrying almost no functional biological information. No gene function, no biosynthetic pathway, no metabolite linkage can be inferred from barcode data alone.
• Tier 0 — No NCBI data: 6,080 endemic species. No sequence data of any kind in any public database.

The combined Tier 0 and Tier 1 population — 9,443 species with either no data or barcode-only data — represents 96% of Madagascar's endemic flora. It is, in practical terms, the complete uncharacterised library. The 3 chromosome-scale genomes and 18 draft genomes represent the narrow sliver that modern genomic science has so far reached.

Why the ratios interact

The three numbers — 83% endemism ratio, 0.72% genomic coverage, 6,080 species with no NCBI data — combine to produce a proposition that is structurally unlike any other geographic opportunity in life sciences data.

The 83% ratio means the biology is non-substitutable: you cannot find these species anywhere else, and there is no surrogate source for their chemistry. The 0.72% coverage means the scientific infrastructure for working with this biology at scale does not yet exist. The 6,080 uncharacterised endemic species means the scale of the unknown is not marginal — it is almost the entire flora.

For pharmaceutical teams, this means that the compound classes available from Malagasy endemics are, with very few exceptions, absent from ChEMBL, PubChem, and every other compound library built from genomic or metabolomic data. The structural diversity available from these 6,080 uncharacterised species represents a non-redundant expansion of accessible chemical space that cannot be approximated by any in-silico approach alone.

For agritech programmes working on climate adaptation traits, the same logic applies. Madagascar's spiny desert and limestone karst floras have evolved drought, heat, saline, and heavy-metal tolerance gene families under conditions far more extreme than those faced by any major crop. Those gene families are, for the most part, in Tier 0 — no sequence, no annotation, no pathway characterisation anywhere in any database.

For AI platforms building biological foundation models, the genomic coverage gap is the training data gap. The models that currently exist have been trained predominantly on the 0.76% of the world's flora that has been sequenced — and predominantly on the crop and model organism species that account for most of that 0.76%. The remaining 99.24% of plant biology, including Madagascar's 9,829 endemics, is invisible to those models. Novel, multi-modal, provenance-clean data from Malagasy endemics is not a marginal supplement to existing training corpora. It is a categorically different type of input.

Data sources and methodology

All endemic counts in this article are derived from the Kew World Checklist of Vascular Plants (WCVP), January 2026 release, covering 362,739 accepted vascular plant species. The WCVP uses a strict native-range filter: only species classified as native, non-extinct, and non-introduced in a given territory are counted. Species with uncertain or transitional native status are excluded. This is a conservative count; the true endemic flora of Madagascar may be larger as recently described species await WCVP review.

Genomic coverage data is from NCBI GenBank's plant assembly summary, January 2026. The "any assembly" count (85 species for Madagascar) includes all assembly types from chromosome-scale to draft; it does not require a complete or high-quality assembly. The chromosome-scale count (3 endemic species for Madagascar) applies a Merqury QV ≥40 / BUSCO ≥90% quality threshold consistent with Earth BioGenome Project standards.

Land area figures are from the World Bank 2022 dataset. Endemic density calculations use the formula: (endemic species count / land area km²) × 1,000.