Human Genetic Variation

Overview: The Scale of Human Genetic Variation

Any two randomly chosen humans differ at approximately 0.1% of their genomic sequence — roughly 3 million positions in the 3.2 billion base pair genome. While this seems small, it generates enormous phenotypic diversity and disease susceptibility differences. Understanding the types and scale of genetic variation is fundamental to human genetics, disease biology, and precision medicine.

Genetic variants are classified by their size and mechanism. Most common is the single nucleotide variant (SNV); when found at a frequency above 1% in a population, it is called a single nucleotide polymorphism (SNP). Larger variants include copy number variants, structural variants, and chromosomal abnormalities.

Single Nucleotide Polymorphisms (SNPs)

Scale of SNP Variation

The human genome contains approximately 4–7 million SNPs per individual relative to the reference genome. The NCBI dbSNP database catalogs over 700 billion known human variants. The gnomAD database (v4.0) contains over 800 million variants observed in ~730,000 sequenced individuals.

Types of SNPs by Genomic Location and Functional Effect

Linkage Disequilibrium and Haplotypes

Nearby SNPs are often inherited together as haplotype blocks due to linkage disequilibrium (LD) — the non-random association of alleles at different loci. LD arises because recombination between close markers is rare. A single tag SNP within an LD block can represent all variants in that region, enabling efficient genotyping for association studies. The HapMap project (2005) and 1000 Genomes project (2015) systematically characterized LD structure and haplotypes across human populations.

Population Differences

SNP allele frequencies vary among human populations as a consequence of demographic history (population bottlenecks, founder effects, migration) and natural selection. African populations have the highest genetic diversity (more SNPs, lower LD blocks) reflecting the origin of Homo sapiens in Africa. Non-African populations show lower diversity reflecting the out-of-Africa bottleneck ~60,000 years ago. These population differences affect polygenic risk score transferability between ancestral groups — a major equity challenge in genomic medicine.

Copy Number Variations (CNVs)

Scale and Discovery

Each person carries approximately 1,000 CNV regions that differ from the reference genome. CNVs collectively account for more base pair differences between individuals than all SNPs combined, yet affect far fewer genomic positions. Landmark studies by Iafrate et al. and Sebat et al. (2004) discovered that large structural variation is pervasive in the human genome — contrary to the prevailing view at the time that the genome was nearly identical in structure between individuals.

CNV Detection Methods

• Chromosomal Microarrays (SNP arrays / CGH arrays): Standard clinical tool; detects CNVs ≥50–500 kb depending on probe density
• Read Depth Analysis from WGS: Detects CNVs as regions of abnormal read coverage (duplications = elevated coverage; deletions = reduced coverage)
• FISH (Fluorescence In Situ Hybridization): Detects specific known CNVs at individual loci using fluorescent probes
• qPCR: Quantifies copy number at specific loci

CNVs and Disease

Some CNV regions are highly associated with disease:

• 22q11.2 deletion syndrome (DiGeorge syndrome): ~3 Mb deletion; most common chromosomal microdeletion (~1/4,000 births); congenital heart defects, cleft palate, T cell immunodeficiency, intellectual disability, schizophrenia risk
• 15q11-q13 duplications/deletions: Angelman syndrome (maternal deletion or paternal UPD) / Prader-Willi syndrome (paternal deletion or maternal UPD) — imprinting-dependent
• SMN1 deletion: Spinal muscular atrophy (SMA) — deletion of SMN1 on chromosome 5q
• Williams syndrome: ~1.8 Mb deletion at 7q11.23 including ELN (elastin)
• Charcot-Marie-Tooth type 1A: 1.5 Mb duplication at 17p11.2 including PMP22
• Autism spectrum disorder: Multiple rare CNVs (16p11.2, 15q13.3, 1q21.1) contribute to ASD risk

Insertions and Deletions (Indels)

Types and Functional Consequences

Frameshift Mutations

When indels in coding regions are not a multiple of 3, they shift the reading frame of translation — changing all subsequent amino acids and usually generating a premature stop codon. Frameshift mutations are typically loss-of-function. Examples:

In-Frame Indels

Indels that are multiples of 3 bp add or remove whole amino acids without shifting the reading frame. These can be pathogenic or neutral depending on the affected protein domain. Example: a 3 bp deletion removing a critical residue in an enzyme's active site.

Examples of Disease-Causing Indels

Microsatellite Instability (MSI)

Microsatellites (short tandem repeats, STRs) are particularly prone to indel mutations due to slippage during DNA replication. Deficiency in DNA mismatch repair (MMR) proteins (MLH1, MSH2, MSH6, PMS2) leads to microsatellite instability (MSI) — widespread accumulation of indels at STR loci throughout the genome. MSI-high cancers (colorectal, endometrial, gastric) have high mutation burdens and are particularly responsive to immune checkpoint inhibitor immunotherapy.

Chromosomal Rearrangements

Large-scale structural changes affect the organization of chromosomal segments and are a major class of human genetic variation, particularly in cancer and developmental disorders.

Types of Chromosomal Rearrangements

Chromothripsis and Chromoplexy

Chromothripsis ("chromosome shattering") is a catastrophic event in which a chromosome undergoes dozens to hundreds of simultaneous double-strand breaks and is reassembled in a disorganized fashion through aberrant DNA repair. First described in 2011, chromothripsis is observed in ~2–3% of cancers (more common in bone cancers) and some congenital developmental disorders. Similarly, chromoplexy involves coordinated rearrangements of multiple chromosomes in a single event — particularly common in prostate cancer.

GWAS — Genome-Wide Association Studies

Genome-wide association studies (GWAS) systematically scan the entire genome to identify common genetic variants (typically SNPs) associated with complex traits and diseases. GWAS transformed human genetics by enabling unbiased discovery of disease-associated loci without prior biological hypotheses.

How GWAS Works

1. Cohort assembly: Cases (individuals with disease) and controls (without disease) are recruited — typically thousands to hundreds of thousands of individuals
2. Genotyping: DNA from each participant is genotyped for 500,000–5 million SNPs using SNP arrays (imputation expands coverage to ~10–30 million common variants)
3. Association testing: For each SNP, a statistical test (logistic regression for case-control, linear regression for quantitative traits) tests whether the SNP allele frequency differs between cases and controls
4. Multiple testing correction: The genome-wide significance threshold is P < 5 × 10⁻⁸ (Bonferroni correction for ~1 million independent tests)
5. Replication: Nominally significant associations are replicated in independent cohorts
6. Fine-mapping and functional annotation: Identify the likely causal variant(s) and mechanism

GWAS Discoveries

The NHGRI-EBI GWAS Catalog contains over 500,000 significant associations across ~4,000 traits as of 2026. Key examples:

Limitations of GWAS

• Missing heritability: Even thousands of GWAS variants explain only a fraction of the genetic contribution to common diseases — rare variants, gene-gene interactions, and epigenetic effects are poorly captured
• Population bias: ~80% of GWAS participants are of European ancestry; PRS and association results often do not transfer well to non-European populations
• Correlation, not causation: GWAS identifies associated SNPs, not necessarily causal variants — fine-mapping and functional studies are required
• Effect sizes are small: Most common variants have odds ratios of 1.05–1.3; not individually clinically actionable

Polygenic Risk Scores (PRS)

Polygenic risk scores aggregate the effects of hundreds to millions of common variants into a single score predicting individual disease risk. PRS for coronary artery disease, breast cancer, and diabetes can identify individuals at 3–5× higher lifetime risk — comparable in risk magnitude to some monogenic risk factors. PRS are beginning to enter clinical practice for coronary artery disease screening in young adults.

Major Human Genetic Variant Databases