Science1 distinct publisher2 min readPublished
In some specimens fewer than 1% of the genome could be read, coverage most pipelines throw away. The Yale team placed those fragments against a tree of good genomes instead, which is why the museum drawer is the real finding.
The Scientist · Science desk
Compiled by The ScientistSomething wrong?How this is made
Phylogenetic placement asks a smaller question than genome assembly, which is why it survives coverage this bad. In some specimens more than 99% of the genome went unread [14], so there is no assembling anything and no calling most sites. A fragment that maps to a known position is still a weak vote about which branch of an existing tree the animal sits on, and the team built that tree first, out of high-quality genomes from living and better-preserved historical tortoises, then placed the scraps against it [4]. The signal comes as much from the reference panel as from the bone.
That is also where the method's reach stops. Caccone's lab has run a long program on the genetic diversity of Galapagos giant tortoises [16], so the comparative material existed before the degraded samples needed it, and the collaborator list runs to nine institutions besides Yale [15], which is a fair measure of what assembling that kind of panel costs in people. A drawer of unidentified bone from a clade nobody has sequenced well has nothing to be placed against. Ochoa's suggestion that the tools transfer to broader conservation work [11] is a claim about the software, not about the reference genomes it assumes are already on hand.
Ochoa named contamination alongside degradation and fragmentation as the problem the study had to work around [8]. The summary does not say how the pipeline told tortoise reads from everything else that has found its way into a museum bone over a century, and that is the section anyone reusing this will turn to first. Two lineages were called from five specimens of extinct tortoises [3]. The per-specimen confidence lives in the paper.
The conservation argument is a chain, and each link is reasonable on its own. Sailors moved tortoises between islands as food, some displaced animals bred with locals, and the genomes of extinct lineages may therefore persist in hybrids alive now [10]; Ochoa suggests captive breeding that mates those hybrids could recover extinct lineage genomes in later generations [11]. What that would produce is a population enriched for ancestral variants, not the Santa Fe tortoise returned. The work reconstructed relationships among lineages [1]. It does not tell you which of those ancestral variants ever did anything for the animal carrying them, and a breeding program picking pairs on placement scores is selecting on ancestry, not on function.
Ranked by verification strength, evidence, and original report placement.
The approach showed that museum specimens previously considered too genetically degraded to be useful in this type of research can help reveal evolutionary history.
Yale-led researchers used a new combination of computational tools to analyze ancient DNA and discovered two ancient, extinct giant tortoise lineages that once lived on the Galapagos Islands; the work reconstructed the genetic relationships of extinct Galapagos giant tortoises.
The study appears in Proceedings of the Royal Society B: Alexander Ochoa et al, 'Integration of ultra-low coverage whole-genome sequences for reconstructing the evolutionary history of Galapagos giant tortoises' (2026), DOI 10.1098/rspb.2026.0103.
The researchers extracted DNA from dried bones of historical museum specimens, including five specimens of two extinct tortoise lineages.
The team combined the scraps of recovered genetic information and superimposed it onto a reference phylogenetic tree built with high-quality genomes from living and other historical tortoises.
In some of the degraded tortoise specimens studied, less than 1% of the genome could be read.
Distinct publishers with included, body-backed reporting in this cluster.
phys.org
1 article · August 27, 2026
Follow any of these and your For You feed starts watching them — no settings page required.
build
The pasted answer transfers the information and skips the judgment rehearsal1 distinct publisher
science
Yale fieldwork finds each Lac Télé village keeping its own flood calendar1 distinct publisher
build
SiFive's BigSky is a porting rig with a rack mount, and that is the honest read on RISC-V servers1 distinct publisher
science
A century of malaria surveys catches warming pushing West Africa past the transmission ceiling1 distinct publisher
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Peer-reviewed anchor, single announcement source, no validation figures
The central result sits in a named, DOI-bearing Proceedings of the Royal Society B paper with a ten-institution author base, which is stronger than a bare vendor or lab claim. But everything reported here comes from one institutional announcement republished by one outlet: the source names no software, releases no accuracy or uncertainty statistics for placing sub-1% coverage fragments, describes no contamination controls, and includes no outside assessment. The method claim is therefore credible but unaudited at this coverage.
One published study, one lab, no external reuse yet
Observed adoption is limited to the originating work: the paper's publication and its single disclosed application to five museum specimens from two extinct lineages within the Caccone lab's ongoing initiative. No other group, collection, breeding program, or software distribution is reported as using the approach, and the broader conservation reuse remains an author projection.
Method reporting restrained; extinct-genome recovery framing runs ahead of data
Most of the article is proportionate: the coverage limit is quantified, the discarded-sample problem is stated plainly, and the placement approach is described without superlatives. The overstatement is concentrated at the ends, where 'discovered two extinct lineages' compresses a phylogenetic placement inference into a discovery, and where recovering extinct-lineage genomes through hybrid captive breeding plus broad wildlife-conservation reuse is offered with no supporting results. Absent validation statistics or any external reuse, the forward claims sit modestly above the evidence.
Institutional research promotion, no counterparty voice
The single source is a university-style research announcement carried by an aggregator: all quotes come from the lead author, the affiliated lab's ongoing initiative is named, and the collaborating institutions receive credit. That is a straightforward reputational and grant-visibility incentive to frame the method as broadly enabling. No commercial product, pricing, or licensing interest is disclosed, and no funder is named, which caps the score below the level of a vendor pitch, but no skeptical or independent voice appears anywhere in the coverage.
Moderate: credible peer-reviewed anchor, single-publisher coverage
Confidence is held mid-range. On the positive side the factual spine is specific and checkable, with a DOI, named journal, named lineages, a quantified coverage figure, and a named author roster. Against that, the cluster contains exactly one publisher, the piece is announcement-derived, and the interpretive claims about generalizability and future genome recovery cannot be tested from what is supplied, so the assessment of method strength and downstream impact remains provisional.