Rare variants are DNA changes carried by fewer than one percent of people in a given population, and they matter because they can have surprisingly large effects on health, disease risk, and drug response. Most of the genetic differences that show up when scientists sequence your genome are common ones shared across large swaths of the population. But over the past decade, large-scale sequencing projects have revealed that each person also carries thousands of uncommon or nearly unique mutations, many of which sit in genes directly relevant to disease. Understanding these variants is reshaping how researchers think about inherited risk, why some people react badly to medications, and how new drugs get developed.
How Rare Variants Arise
Every time a cell copies its DNA, there is a small chance of introducing a new mutation. Most of the time these changes are harmless, but occasionally they land in a gene that matters. If a mutation is harmful enough, natural selection tends to weed it out before it spreads widely, which is precisely why damaging variants stay rare. But they never disappear entirely, because new ones keep arriving with each generation.
The explosive growth of the human population over the last several thousand years has amplified this effect dramatically. As a population balloons, more babies are born, which means more new mutations enter the gene pool. Many of these are too recent to have been filtered out by natural selection yet. Simulations show that population growth increases the number of mildly harmful mutations each person carries, but the average impact of each individual mutation becomes somewhat less severe because selection gets better at removing the worst offenders while milder ones slip through.1PubMed Central. Population growth inflates the per-individual number of deleterious mutations and reduces their mean effect The practical result is that the genetic risk for common diseases in modern populations is spread across a larger number of weakly damaging rare variants rather than concentrated in a few strongly damaging ones.
A sequencing study of more than 14,000 people across 202 drug-target genes confirmed this picture directly. The researchers found an abundance of rare functional variants and concluded that rapid population growth, combined with weak purifying selection, accounts for why we carry so many of them.2PubMed Central. An abundance of rare functional variants in 202 drug target genes sequenced in 14,002 people Many of these variants are so new to the population that they appear in only a single individual within a study sample. Population growth is predicted to greatly increase the burden of these private mutations for each person.3PubMed Central. High burden of private mutations due to explosive human population growth and purifying selection
Bigger Effects Than Common Variants
One of the most striking findings from recent genetics research is the gap in punch between rare and common variants. Common variants, the kind cataloged in large genome-wide association studies, typically nudge disease risk by small amounts. Rare variants in the same genes can hit much harder. An analysis of nearly 455,000 exomes found that rare, penetrant mutations in genes identified by genome-wide association studies confer roughly ten-fold larger effects than common variants in those same genes.4PubMed. Rare penetrant mutations confer severe risk of common diseases In other words, if a common variant in a gene raises your risk of a disease by a few percentage points, a rare variant in the same gene might multiply your risk several times over.
This asymmetry exists because natural selection keeps a tight lid on anything too dangerous. A variant that dramatically raises the risk of dying young or not reproducing almost never reaches high frequency. It either gets eliminated quickly or lingers at the fringes of the population. Common variants, by contrast, have survived generations of selection, which generally means their effects are mild enough to be tolerated. The trade-off is that when you do carry a rare damaging variant, it tends to pack a bigger punch precisely because selection has not had the chance to remove it yet.
Solving the Missing Heritability Puzzle
For years, geneticists faced a frustrating problem. Twin and family studies suggested that many common diseases and traits are highly heritable, yet the common variants identified in large association studies only explained a fraction of that inherited risk. This gap became known as “missing heritability.” Part of the answer, it turns out, lies in rare variants.
A whole-genome sequencing analysis found that rare variants, especially those in regions of the genome where they are not easily detected by standard genotyping chips, are enriched for heritability. The enrichment was even stronger for variants that alter protein structure, consistent with what you would expect if negative selection were keeping these changes rare. The researchers concluded that rare variants are a major source of the still-missing heritability of complex traits and disease.5PubMed Central. Assessing the contribution of rare variants to complex trait heritability from whole genome sequence data This does not mean rare variants explain everything, but they fill a substantial part of the gap that common variants alone could not account for.
Why Finding Them Is Hard
If rare variants matter so much, why has it taken so long to study them? Two major hurdles stand in the way: detection and statistics.
On the detection side, older genotyping technology relied on microarrays that measure a pre-selected set of known common variants. By design, these chips miss variants that are new or rare. Deep whole-genome sequencing, reading every letter of DNA at high coverage, changed the game. A study that sequenced more than a thousand Japanese individuals at high depth demonstrated that this approach could identify very rare and novel variants that were impossible to find with microarrays and difficult to catch with lower-coverage sequencing.6Nature Communications. Rare variant discovery by deep whole-genome sequencing of 1,070 Japanese individuals The cost of sequencing has dropped fast enough that biobanks now routinely sequence hundreds of thousands of participants, making rare variant discovery increasingly practical.
The statistical challenge is different but equally stubborn. Standard association tests compare how often a variant appears in people with a disease versus those without. When a variant is carried by only a handful of individuals, there simply is not enough data to tell whether the association with disease is real or just a fluke. Researchers developed gene-based tests that group all rare variants within a gene and test them collectively. One widely used method, the sequence kernel association test, treats rare variants as a set and asks whether the combined signal across the gene is stronger than chance would predict.7PubMed Central. Rare-variant association testing for sequencing data with the sequence kernel association test A later extension, SKAT-O, improved on this by adaptively deciding whether to weight all variants in the same direction or allow some to have opposing effects, depending on what the data suggest.8PubMed Central. Optimal unified approach for rare-variant association testing with application to small-sample case-control whole-exome sequencing studies These methods opened the door to finding meaningful rare variant associations that one-at-a-time testing would have missed.
The Problem of Uncertain Significance
Even after you find a rare variant, figuring out whether it actually does anything is a separate headache. Clinical genetics labs classify variants on a five-tier scale from “benign” to “pathogenic.” A huge number of rare variants land in the uncomfortable middle category: variants of uncertain significance, or VUS. Only a minority of VUS are later reclassified as pathogenic when new evidence emerges, and that reclassification process is rarely timely. In the meantime, the uncertainty complicates clinical decisions, can lead to unnecessary treatment, and causes real anxiety for patients.9PubMed Central. The Challenge of Genetic Variants of Uncertain Clinical Significance : A Narrative Review
The problem is particularly acute in genes with serious clinical implications. In TP53, a gene whose mutations are linked to a hereditary cancer syndrome, the growing use of multigene panels has led to a flood of VUS results. When there is no clear family history of the associated syndrome, both patients and clinicians are left in a difficult position, unable to dismiss the finding or act confidently on it.10PubMed. TP53 variants of uncertain significance: increasing challenges in variant interpretation and genetic counseling
Computational tools try to bridge this gap by predicting whether a rare variant is likely to be damaging. A recent benchmark of 28 prediction methods found that the best-performing tools combined evolutionary conservation scores, predictions from other algorithms, and allele frequency data to achieve the highest accuracy. But a consistent pattern emerged: as variants became rarer, prediction specificity dropped, meaning the tools increasingly flagged harmless variants as harmful.11PubMed Central. Assessing the performance of 28 pathogenicity prediction methods on rare single nucleotide variants in coding regions Ensemble methods that integrate many prediction scores into a single number, like REVEL, perform better than any individual component, but even the best still make mistakes.12American Journal of Human Genetics. A Rare Exome Variant Ensemble Learner to Improve Pathogenicity Prediction for Imbalanced Data The gap between finding a rare variant and knowing what it means for a patient remains one of genomics’ most pressing problems.
How Rare Variants Affect Drug Response
Your genes influence how your body processes medications, and rare variants in drug-metabolizing enzymes can be the difference between a drug working normally and causing a dangerous reaction. Genetic variability in enzymes, transporters, and receptors involved in drug metabolism shapes how you absorb, distribute, break down, and excrete medications.13PubMed Central. Integrating rare genetic variants into pharmacogenetic drug response predictions
Most pharmacogenomic testing today focuses on a handful of well-studied common variants, but rare variants can cause the same kinds of problems. In a study of people who developed a dangerous heart rhythm abnormality called acquired long QT syndrome after taking medication, about eight percent carried a rare variant in a cytochrome P450 gene. Of the identified variants, nearly all were in enzymes known to metabolize at least one of the drugs the patient had been taking. The study found an increased burden of rare variants in these drug-processing genes compared to controls.14PubMed. Rare Variation in Drug Metabolism and Long QT Genes and the Genetic Susceptibility to Acquired Long QT Syndrome
A similar story plays out with fluoropyrimidine chemotherapy drugs. The enzyme DPD, encoded by the gene DPYD, breaks down these drugs. Patients with low DPD activity can experience severe toxicity. Researchers found that rare variants in DPYD were significantly more common in patients with low DPD activity, appearing in about nine percent of low-activity patients compared to about three percent of those with normal activity. Even after excluding the rare variants already recognized as clinically relevant, the enrichment held.15The Pharmacogenomics Journal. Integrating rare genetic variants into DPYD pharmacogenetic testing may help preventing fluoropyrimidine-induced toxicity Standard testing panels would miss these variants entirely, leaving some patients exposed to preventable harm.
Cancer Genomics and Rare Somatic Mutations
Rare variants are not just inherited. Tumors accumulate mutations as they grow, and many of these somatic variants are rare in the sense that each one appears in only a tiny fraction of cancer patients. The challenge is figuring out which rare mutations are actually driving the cancer and which are harmless passengers.
Standard methods for identifying cancer driver genes rely on how frequently a gene is mutated across tumor samples. This approach works well for obvious hotspots but misses the long tail of rare mutations that may be doing real damage. A protein domain-based framework called Oncodomains addressed this by comparing mutations across families of related genes. The variants it identified occurred in an average of just over one tumor sample each, far below the threshold that frequency-based methods could detect, yet they fell in protein regions with clear functional relevance.16PLoS Computational Biology. Oncodomains: A protein domain-centric framework for analyzing rare variants in tumor samples
A pan-cancer analysis took a related approach, looking at mutations scattered across large gene families. It found that rare, individually infrequent mutations in upstream signaling genes, including hundreds of G protein-coupled receptor genes, showed patterns of mutual exclusivity with known oncogenic hotspots. That mutual exclusivity is a strong hint that the rare mutations are doing the same job as the famous driver mutations, just in different patients.17Oncogene. Rare, functional, somatic variants in gene families linked to cancer genes: GPCR signaling as a paradigm
Natural Knockouts as a Window Into Drug Targets
Some rare variants completely knock out a gene’s function, providing a natural experiment in what happens when that gene is turned off in a living person. These loss-of-function variants are extraordinarily useful for drug development. If you are considering developing a drug that inhibits a particular protein, you can look at people who carry natural loss-of-function variants in the gene encoding that protein and see what happens to them. Healthy carriers with no obvious problems suggest the drug might be safe. Carriers with specific health benefits point toward a promising target.
Naturally occurring loss-of-function variants provide an in vivo model of gene inactivation that complements laboratory knockout experiments in cells and animal models. An analysis from a large exome sequencing consortium reported that even essential genes, where complete loss of function is not tolerated, can be highly successful as targets for inhibitory drugs.18PubMed Central. Evaluating drug targets through human loss-of-function genetic variation This finding matters because it means researchers do not have to rule out a drug target simply because the underlying gene appears essential.
Large biobanks have accelerated this work. An analysis of over 400,000 UK Biobank participants identified 18 new associations between rare predicted loss-of-function variants and metabolic or cardiometabolic traits.19Nature Communications. Analysis of predicted loss-of-function variants in UK Biobank identifies variants protective for disease A parallel study of more than 337,000 individuals in the same biobank found 291 genes with homozygous loss-of-function carriers that had never been observed before, expanding the catalog of human genes for which complete knockouts exist in real people.20Nature Communications. Medical relevance of protein-truncating variants across 337,205 individuals in the UK Biobank study Each of those carriers is essentially a walking experiment in what that gene does in a human body.
Carrier Effects in Genes Linked to Severe Childhood Disorders
An intriguing finding from large-biobank research is that carrying just one copy of a variant associated with a severe recessive disorder, a situation where you would not develop the disorder itself, can still have measurable consequences. An analysis of UK Biobank participants who carried rare, damaging variants in genes known to cause developmental disorders found a pattern of mildly impaired cognition, shorter stature, higher body mass index, and socioeconomic disadvantage compared to non-carriers. These individuals were also more likely to have been diagnosed with a psychiatric condition or childhood developmental disorder, and men with these variants had fewer children.21PubMed Central. A spectrum of recessiveness among Mendelian disease variants in UK Biobank The effects were subtle compared to what a child with two copies of such a variant would experience, but they were statistically clear across a large population. This challenges the classic textbook notion that recessive carriers are unaffected, and it has implications for understanding why certain harmful variants persist in the population.
Beyond Protein-Coding Genes
The vast majority of the human genome does not code for proteins. Instead it contains regulatory elements, switches that control when and where genes are turned on or off. Rare variants in these non-coding regions represent a vast frontier that is far less explored than the protein-coding exome, partly because it is harder to predict what a change in a regulatory region will do. Yet some research suggests non-coding regions could play a major role in disease susceptibility.22PubMed. Rare variant association testing in the non-coding genome
Early results hint that rare variants in regulatory DNA behave differently from common ones. A preprint mapping regulatory effects in brain and heart tissue found that while common variant effects tend to be specific to particular cell types, rare variants exert broader regulatory effects shared across cell types, with natural selection especially targeting variants that affect fetal brain neurons.23bioRxiv. Mapping the regulatory effects of common and rare non-coding variants across cellular and developmental contexts in the brain and heart If confirmed, this would suggest that rare regulatory mutations can have more pervasive biological effects than their common counterparts, even outside of genes themselves. The field is still in early stages here, but non-coding rare variants are widely expected to explain a significant additional slice of disease risk as sequencing studies scale up.
Structural variants add another layer. Rather than single-letter changes, these involve deletions, duplications, or rearrangements of larger chunks of DNA. A study using UK Biobank data found that deletion of an entire enhancer element had roughly a five-fold larger effect on platelet counts than single-letter changes within that same enhancer, illustrating that structural variants can alter the genome in ways that single-nucleotide changes cannot replicate.24Cell. Haplotype-informed structural variant calling and association analysis in UK Biobank
Why Population Diversity in Databases Matters
Whether a variant counts as “rare” depends on who you are comparing it to. A variant might be rare globally but relatively common in a specific population, or vice versa. Most large genomic reference databases have been built primarily from people of European descent, which creates a serious blind spot. When a clinician looks up a variant and finds no population frequency data because the patient’s ancestry is underrepresented in the database, the variant is more likely to be classified as uncertain.
This is not a hypothetical problem. A study comparing variant frequencies between a Brazilian cohort and a widely used global database found that more than 40 percent of shared variants had significantly different frequencies between the two. Among variants classified as uncertain significance, 20 had high enough frequency in the Brazilian population to be reclassified as likely benign, but they remained flagged as uncertain in other populations because the data to make that call simply did not exist.25PubMed. Variant classification of hereditary cancer genes is affected by genomic underrepresentation of admixed populations Similarly, interpreting BRCA1 variants in Middle Eastern populations is complicated by the scarcity of ancestry-matched reference data, which can lead to overestimating how rare a variant actually is.26Biomolecules and Biomedicine. Reclassification of BRCA1 variants of uncertain significance using ancestry-stratified allele frequencies: A Middle East-focused analysis
The rare variant spectrum also differs between populations at a fundamental mutational level. An analysis across global populations found distinct mutational signatures in rare variants that track with geography: one signature characterized by specific mutation types almost perfectly separates Western Eurasians from other groups, while another is restricted to certain South and Central American populations but absent from closely related Amazonian groups.27PLoS Genetics. Differences in the rare variant spectrum among human populations These differences are not just academic curiosities. They mean that prediction tools, risk calculators, and clinical classification systems trained primarily on European data will perform less well for everyone else, with real consequences for diagnosis and treatment. Expanding the diversity of reference databases is one of the most straightforward ways to improve rare variant interpretation for all patients.