Variant interpretation has advanced rapidly over the past decade, yet one fundamental dimension of biological evidence remains largely absent from clinical and research workflows. That missing dimension is evolutionary time.
Most current interpretation frameworks rely on evidence derived almost exclusively from human data. Allele frequencies are estimated from large population cohorts. Clinical assertions are drawn from case reports and curated databases. Functional assays interrogate the effect of specific variants in controlled experimental systems. Each of these evidence types is valuable, but together they still leave a large fraction of human variation unresolved.
The persistence of variants of uncertain significance is not a failure of effort or data collection. It reflects a structural limitation in how evidence is framed. Human-only data, no matter how large, capture only a narrow window of the evolutionary processes that define gene function.
The blind spot created by human-only evidence
Human genetic diversity is shallow in evolutionary terms. Modern humans share a recent common ancestry, and differences among individuals reflect tens of thousands of years of divergence, not the millions of years over which genes and proteins evolved their functional constraints. As a result, most positions in the genome have not been meaningfully tested by natural selection within the human lineage alone.
This limitation becomes apparent in variant interpretation. Rare missense variants dominate clinical sequencing results, yet most occur at positions for which human population data provide little discriminatory power. Functional assays are unavailable for the vast majority of these variants, and in-silico predictions frequently disagree. The outcome is a large residual class of variants that cannot be confidently classified.
From Cornerstone Genomics’s perspective, this uncertainty arises because current workflows ask the wrong question. Instead of asking only whether a variant is rare in humans or previously observed in patients, we must also ask whether the underlying position in the gene has ever tolerated change across evolutionary history.
What evolution reveals that humans alone cannot
Evolutionary constraint captures the outcome of millions of years of natural selection acting on genes and proteins. Positions that are essential for function tend to remain conserved across species. Positions that are more flexible accumulate variation without compromising viability. These patterns are not theoretical. They are directly observable when sequences are compared across appropriate evolutionary distances.
Large comparative genomics studies have demonstrated that conserved genomic regions are enriched for functional elements and disease-associated variation, confirming that evolutionary signatures correlate with biological importance (Goode et al. 2010; Lindblad-Toh et al. 2011). More recent work across expanded mammalian and primate datasets has further refined these constraint maps and shown that evolutionary conservation provides predictive power beyond human-only data (Kuderna et al. 2024).
What is often underappreciated is that the choice of species matters. Comparisons that are too distant collapse most of the genome into uniformly conserved regions. Comparisons that are too close, such as human-to-human or human-to-great-ape alone, fail to reveal which changes are naturally tolerated. The evolutionary window must be wide enough to sample variation, but close enough to preserve functional relevance.
Why primates matter for medical genomics
Cornerstone Genomics approaches this problem from the perspective of primate evolutionary biology. Humans are one genus within the primate order. Across approximately 80 million years, primates diversified into dozens of lineages, each carrying the same genes but different versions shaped by distinct evolutionary pressures.
By examining how protein-coding genes vary across primates, it becomes possible to observe which amino acid changes persist in nature and which do not. This distinction is critical for variant interpretation. Variants that recur across primate lineages have passed repeated functional tests imposed by natural selection. Variants that are unique to humans, particularly at highly constrained positions, warrant closer scrutiny.
Cornerstone Genomics’s proprietary primate exome datasets were generated specifically to capture this evolutionary signal. By sequencing and aligning coding regions from 55 primate genera and anchoring them to the human reference genome, Cornerstone Genomics created a resource that directly links evolutionary history to human variant interpretation. This work was not motivated by clinical annotation gaps, but by a long-standing question in evolutionary genomics: how much change can each gene tolerate while remaining functional.
Evolution as an orthogonal evidence axis
Evolutionary evidence does not replace population frequency, clinical assertions, or functional assays. It complements them. Allele frequency tells us how common a variant is today. Clinical data tell us what has been observed in patients. Functional assays test specific mechanisms. Evolutionary constraint tells us what biology has already tested across millions of years.
This evidence axis is orthogonal because it is independent of human sampling bias, cohort size, and clinical ascertainment. It provides information even when a variant has never been seen before in humans. That property makes it uniquely valuable for interpreting rare and novel variants, which now dominate sequencing output.
From evolutionary insight to practical interpretation
CodeXome operationalizes this evolutionary perspective by integrating primate comparative genomics with existing human-derived evidence. Variants are evaluated in the context of deep evolutionary tolerance. Positions that show extensive natural variation across primates are deprioritized, while variants at evolutionarily constrained sites are highlighted for further review.
In validation studies using ClinGen and BRCA Exchange datasets, variants classified as benign or likely benign were consistently shared across primates, while pathogenic variants were absent from natural primate variation. Importantly, a substantial fraction of variants previously classified as uncertain could be clarified based on their evolutionary context. These observations reflect a biological principle rather than a statistical artifact: deleterious variants do not persist in nature.
Implications for variant interpretation workflows
Incorporating evolutionary evidence changes how effort is allocated. Instead of manually reviewing thousands of variants with limited context, interpretation can focus on a smaller, biologically enriched subset. This shift reduces curation burden, improves consistency, and creates clearer decision points for functional follow-up or clinical reporting.
For clinical laboratories, this means faster triage and fewer unresolved results. For medical geneticists, it provides mechanistic context that human-only data cannot supply. For researchers, it enables more targeted hypothesis generation grounded in biological constraint rather than frequency alone.
Looking forward
Human genomes carry the imprint of deep evolutionary history, and ignoring that history leaves critical information unused. Evolutionary genomics adds a missing dimension that helps explain why some variants matter and others do not.
Cornerstone Genomics was built at the intersection of primate biology, evolutionary genomics, and human disease. CodeXome reflects that origin by bringing evolutionary evidence directly into interpretation workflows. We invite researchers, clinicians, and laboratories to explore how evolutionary filtering can complement existing approaches and to collaborate on applying this perspective to challenging variant interpretation problems.
References
- Goode DL et al. 2010. Evolutionary constraint facilitates interpretation of genetic variation in resequenced human genomes. Genome Research.
- Lindblad-Toh K et al. 2011. A high-resolution map of human evolutionary constraint using 29 mammalian genomes. Nature.
- Kuderna LF et al. 2024. Identification of constrained sequence elements across primates and mammals. Nature.
