Ancient Tooth Proteins Tell Tales

An analysis of six ancient Homo erectus tooth proteins from three locations in China, a Denisovan tooth protein, and some modern human and animal tooth proteins reveal some notable insights. Bernard's Blog has the story. Proteins in teeth are easier to recover than ancient DNA and can serve as a proxy for it in cases like these.

Interestingly, it implies that a tooth enamel protein found in 1% of humans in or near the former Denisovan range, and Denisovans have a tooth enamel protein also found in Chinese Homo erectus, probably due to introgression from Homo erectus to Denisovans and then from Denisovans to modern humans, in the view of the researchers.

John Hawk has a discussion of the paper, however, that interprets the data differently:

All six teeth share two derived amino acid changes, both in the sequence of an enamel matrix protein known as ameloblastin, or AMBN. One of these hasn’t before been seen in hominins: a change from alanine to glycine at position 253 of the sequence, or A253G. The other is 20 positions downstream, swapping in valine for the ancestral methionine, M273V. The DNA mutation encoding this change is shared by both the Denisova 3 and Denisova 25 genomes. The M273V amino acid change itself is in the Harbin and Penghu 1 dental proteomes—part of why they align with Denisovans. A number of genomes from modern people also share this change, possibly from Denisovan ancestors.

The hypothesis presented by Fu and coworkers is that Homo erectus was the source of M273V, and its presence in Denisovans is a result of introgression. In support of this idea of introgression, they note earlier research on the Denisova 3 and Denisova 25 genomes that suggests a contribution from a “superarchaic” source population. Many—including me—have speculated that this superarchaic ancestry came from H. erectus. Fu’s team may have just proved it.

But I don’t think these teeth are Homo erectus.

Their estimated ages, all around 400,000 years ago, are prime Denisovan time. Fu and coworkers find that all the teeth share a derived link with later Denisovan genomes. For me, the most likely hypothesis is that these teeth come from a population within the Denisovan branch of humanity.

This situation is basically the same as the Sima de los Huesos fossils. Those remains are around 430,000 years old. Those fossils look like Neanderthals in some subtle ways, but until DNA was recovered from them, many researchers considered them to be part of a different group, often called Homo heidelbergensis. DNA revised both the timeline and their identification.

It may seem heretical, but I think protein data may be about to do the same for fossils from East Asia.

How Common Are Hallucinated Citations In Academic Papers?

Academic papers written with AI that contain hallucinated citations are on the rise.

Large language models (LLMs) are known to generate plausible but false information across a wide range of contexts, yet the real-world magnitude and consequences of this hallucination problem remain poorly understood. 

Here we leverage a uniquely verifiable object - scientific citations - to audit 111 million references across 2.5 million papers in arXiv, bioRxiv, SSRN, and PubMed Central. We find a sharp rise in non-existent references following widespread LLM adoption, with a conservative estimate of 146,932 hallucinated citations in 2025 alone. 

These errors are diffusely embedded across many papers but especially pronounced in fields with rapid AI uptake, in manuscripts with linguistic signatures of AI-assisted writing, and among small and early-career author teams. At the same time, hallucinated references disproportionately assign credit to already prominent and male scholars, suggesting that LLM-generated errors may reinforce existing inequities in scientific recognition. Preprint moderation and journal publication processes capture only a fraction of these errors, suggesting that the spread of hallucinated content has outpaced existing safeguards. Together, these findings demonstrate that LLM hallucinations are infiltrating knowledge production at scale, threatening both the reliability and equity of future scientific discovery as human and AI systems draw on the existing literature.

Zhenyue Zhao, Yihe Wang, Toby Stuart, Mathijs De Vaan, Paul Ginsparg, Yian Yin, "LLM hallucinations in the wild: Large-scale evidence from non-existent citations" arXiv:2605.07723 (May 8, 2026).

Extended Tully-Fisher Relations

Stacy McGaugh's latest post explores in greater depth a major paper is was a co-author of in March of this year which demonstrates how tightly the baryonic Tully-Fisher relation holds from the smallest galaxies to the largest structures in the universe with very little intrinsic scatter. It also highlights that fact that while this scaling law is displaced in galaxy cluster and larger structures, that the same basic scaling law and slope, albeit displaced, holds there as well, as shown in the second chart below.

My take is that the geometry of the mass distribution or interstellar media in clusters could explain the displacement.