Massively parallel functional screen identifies thousands of regulatory differences in human versus chimpanzee postcranial skeletal development

Every element of the human skeleton exhibits some differences in comparison to our closest living relatives. Many of these skeletal modifications underpin key events in human evolution, enabling our species to walk upright, manipulate tools with precision, and support enlarged brains. Accordingly, many derived human features related to changes in the brain, locomotion, diet, and life history have been under natural selection since our last common ancestor with chimpanzees. Other changes result from random processes such as genetic drift. Identifying the genomic changes that underlie human-specific features of the skeleton (and other tissues) remains an outstanding challenge due to the substantial number of differences between the human and chimpanzee genomes. Approximately 3,000 of these changes are human accelerated regions (HARs) or regions that have accumulated more nucleotide changes than expected by chance, while others are single base-pair substitutions or other genomic signals. For the most part, a comprehensive functional analysis of all such changes as they pertain to the skeleton, let alone any tissue, remains lacking. Therefore, to identify human-chimp sequence differences that modulate gene expression in the developing postcranial skeleton, we used a massively parallel reporter assay (MPRA) to screen the human and chimp versions of 70,000 regulatory elements present in the prenatal cartilage template for differential activity. After testing our library in two cartilage and one bone marrow-derived lymphoblast line, we identify 30,736 regions (45.2%) with activity in our assay. Of the active regions, we find that 11,542 (37.6%; or 17% of the entire pool) regions exhibited differential activity between the human and chimpanzee. We find that human ancestor quickly evolved regions (HAQERs) were predictive of differential activity while HARs were not and both sets failed to predict the magnitude of effect, unlike the total number of base pair differences between species, which was weakly correlated with effect size. These findings reveal that human skeletal evolution involves widespread regulatory changes distributed across thousands of elements rather than concentrated effects at a few key loci, supporting a polygenic model for the evolution of complex morphological traits. This comprehensive dataset provides a large resource for understanding the regulatory basis of human skeletal adaptations and demonstrates the necessity of large-scale functional validation for interpreting evolutionary genomic comparisons. Overall design: MPRA testing the regulatory potential of 70,000 regulatory elements active in the prenatal cartilage template for differential activity between humans and chimpanzees. Five replicates were performed in K562, TC28a2, and CHON002 cells.