Human joints carry a quarter to a third of the glycosaminoglycan found in other apes

Each candidate element in the screen sat upstream of a transcribable DNA barcode, and the RNA abundance of each barcode is the readout for what that sequence drives ^[8]. The human allele and the great ape allele of every element went into the same pool of 732,804 sequences, so both met the same cells under the same conditions ^[7]. The cells were chondrocytes, picked because they form the skeletal scaffold that most bones develop from and remain the predominant resident cell type in adult cartilage ^[11].

The design rests on a premise the paper states directly: divergence between close relatives is thought to come mainly from gene-regulatory change, and cis-regulatory change in particular is considered a major driver of morphological difference ^[13]. A reporter assay measures the expression a sequence drives from a construct ^[8]. What that element does at its own position in the genome of a growing limb is a separate question.

The two big counts are in different units. The screen covered 561,410 human-derived substitutions ^[1], while 15,077 is a count of loci with human-specific regulatory activity ^[2]. Dividing one by the other gives 2.7 percent, about one per 37 substitutions assayed ^[14]. A single locus can hold more than one substitution, so that figure is not a per-variant hit rate.

For scale, an earlier MPRA by the same authors covered the 14,042 substitutions that separate modern humans from Neanderthals and Denisovans ^[9]. The new screen is roughly 40 times larger by that count ^[15]. Earlier reporter work on the deeper human-great ape divergence went mostly into neural cell types, or into restricted classes of loci such as human accelerated regions ^[10].

Stated the other way round, a three-to-fourfold reduction puts human joint glycosaminoglycan content at a quarter to a third of the level in non-human apes ^[16]. The route to it in the paper is regulatory: the atlases show suppression of GAG biosynthesis inside a broader rewiring of the extracellular matrix ^[5]. The abstract, in Nature, reports the joint content difference; the method of measurement goes unstated ^[17][12].

The claim about disease rests on different evidence from the claim about selection. The authors wrote that the human-specific shift "is likely to be a key contributor to the exceptional susceptibility of humans to degenerative skeletal diseases" ^[6], and they cite three earlier papers for that susceptibility ^[18]. A selection signature tells you a substitution's history looks unlike drift, and leaves open which phenotype selection acted on. The experiments reported here are in cultured chondrocytes and in human-ape hybrid cells differentiated into osteochondral progenitors ^[1][3].