Gene Duplication

Why gene duplication matters

• It is the main engine of new genes. A gene doing an essential job is trapped by purifying selection — almost any mutation that alters it is eliminated. Duplication breaks the trap: one copy keeps the essential job while the spare explores mutation space. Most genes in every sequenced genome belong to multi-member families created this way.
• It built your color vision. Old World primates see in trichromatic color because an ancestral opsin gene on the X chromosome duplicated roughly 30 to 40 million years ago and diverged into the red (OPN1LW) and green (OPN1MW) pigments, whose peak sensitivities differ by only about 30 nanometers. New World monkeys mostly lack this duplication and rely on allelic polymorphism instead.
• It underlies the oxygen-carrying globins. The alpha- and beta-globin clusters, myoglobin, and neuroglobin all descend from one ancestral globin gene through successive duplications. Tandem duplicates within each cluster (embryonic, fetal, and adult chains) let mammals switch hemoglobins across development — fetal HbF has higher oxygen affinity than adult HbA to pull oxygen across the placenta.
• It patterned the animal body plan. The four vertebrate Hox clusters (HoxA–D), which lay out the head-to-tail body axis, arose from a single ancestral cluster amplified by tandem duplication and then multiplied by two rounds of whole-genome duplication (the 2R hypothesis). Amphioxus, a basal chordate, still has just one cluster.
• It fuels adaptation to extremes. Antarctic notothenioid fish survive sub-zero seas because a duplicated trypsinogen (digestive enzyme) gene was co-opted into an antifreeze glycoprotein — a textbook case of neofunctionalization. Duplication also expanded snake and cone-snail venom families, insecticide-resistance esterases, and amylase copy-number in starch-eating human populations.
• It drives disease and cancer. The same misaligned-repeat recombination that creates duplicates also causes genomic disorders: Charcot–Marie–Tooth disease type 1A comes from a 1.4-megabase duplication carrying PMP22. In cancer, amplification of oncogenes such as MYC, ERBB2 (HER2), and EGFR is duplication run amok — and HER2 amplification is the target of trastuzumab.

How gene duplication works

A duplication has two acts: making the extra copy, then deciding its fate. The copy can be produced several ways. Unequal crossing over (non-allelic homologous recombination) happens when repetitive sequences on paired homologous chromosomes misalign during meiosis; a crossover then copies a stretch onto one chromatid while deleting it from the other. This mechanism generates the tandem arrays seen in the globin and Hox clusters, and — running in reverse — the reciprocal deletions that cause disease. Retrotransposition occurs when a mature, spliced mRNA is reverse-transcribed by LINE-1 machinery and reinserted elsewhere; the resulting retrocopy lacks introns and usually a promoter, so it typically arrives dead as a processed pseudogene. Segmental duplication copies genomic blocks larger than a kilobase; these low-copy repeats make up roughly 5% of the human genome and seed recurrent structural variation. And whole-genome duplication (polyploidy) doubles every gene at once.

Once a duplicate exists, its default destiny is death. Because the original still covers the job, the spare is redundant, purifying selection relaxes on it, and mutations accumulate freely. Most duplicates therefore collect a frameshift, a premature stop codon, or a promoter lesion and become a pseudogene — nonfunctionalization. This is the fate of the great majority of duplicates, which is why genomes are littered with 14,000 to 20,000 pseudogenes.

A minority escape decay. In neofunctionalization, the redundant copy — before it is disabled — happens on a beneficial mutation that gives it a genuinely new function; selection then preserves it. In subfunctionalization, the two copies partition the ancestral gene's sub-functions between them so that both become indispensable. The duplication–degeneration–complementation (DDC) model of Force, Lynch, and colleagues (1999) showed this can happen through degenerative mutations alone: if the ancestral gene had two tissue-specific enhancers, one copy can lose enhancer A while the other loses enhancer B, and now neither is dispensable. Because subfunctionalization needs only common loss-of-function mutations — not a rare beneficial one — it can preserve duplicates quickly, buying time in which neofunctionalizing mutations may later arise (the "sub-neofunctionalization" continuum).

The three fates of a duplicate gene

Fate	What happens to the copies	Mutations required	Textbook example
Pseudogenization (nonfunctionalization)	One copy silenced by frameshift, stop codon, or promoter loss	Common loss-of-function	~14,000–20,000 human pseudogenes; olfactory receptor pseudogenes
Neofunctionalization	Old copy keeps ancestral job; spare gains a new function	Rare beneficial (gain-of-function)	Antifreeze glycoprotein from trypsinogen in notothenioid fish
Subfunctionalization (DDC)	Ancestral sub-functions split between the two copies	Complementary loss-of-function	Zebrafish engrailed and many teleost duplicate pairs

Paralogs vs orthologs

Property	Paralogs	Orthologs
Origin of the split	Gene duplication within a genome	Speciation between lineages
Relationship	Homologs that co-exist in the same genome	Homologs in different species
Function	Free to diverge; often gain new roles	Tend to conserve the ancestral function
Example	Human alpha-globin vs beta-globin	Human beta-globin vs mouse beta-globin
Annotation transfer	Risky — divergence common	Reliable — the "ortholog conjecture"
Term coined by	Walter Fitch, 1970

Common misconceptions

• "A duplicate is instantly a new gene." At the moment of duplication the two copies are identical and redundant. Novelty requires subsequent mutation and selection over many generations; the default and most common outcome is not innovation but silent decay into a pseudogene.
• "Neofunctionalization is the main fate." Genuinely new functions are the exception. Pseudogenization is the statistical rule, and among retained duplicates, subfunctionalization is often more common early on because it needs only ordinary degenerative mutations rather than a rare beneficial one.
• "Paralogs and orthologs are the same as homologs." All paralogs and orthologs are homologs, but the reverse is not true. Homology just means shared ancestry; paralog (duplication split) and ortholog (speciation split) specify how two homologs came to be separate.
• "Pseudogenes are junk with no role." Most are inert, but some are transcribed and regulatory. PTENP1, a pseudogene of the tumor suppressor PTEN, produces an RNA that sponges microRNAs, indirectly raising PTEN levels — deletion of PTENP1 is selected for in some cancers.
• "Whole-genome duplication is a plant-only oddity." Polyploidy is spectacularly common in plants (all flowering plants share an ancient event; wheat is hexaploid), but vertebrates carry the signature of two ancient rounds, and teleost fish an additional third — which is why zebrafish frequently have two genes where humans have one.
• "More gene copies is always better." Copy number is under selection in both directions. Extra copies impose a cost (dosage imbalance, misregulation), and gene amplification of oncogenes such as MYC and HER2 is pathological. The dosage-balance hypothesis explains why some gene classes are retained after WGD but lost after small-scale duplication.

Famous experiments and history

• Bridges and the Bar eye (1936). Calvin Bridges showed that the Bar mutation in Drosophila, which narrows the eye, is a tandem duplication of a chromosomal band visible in the polytene chromosomes — the first physical demonstration that genes can be duplicated, decades before DNA sequencing.
• Ingram and hemoglobin fingerprinting (1950s–60s). Vernon Ingram's protein sequencing revealed that the alpha and beta hemoglobin chains, and myoglobin, share deep sequence similarity, implying they descend from one ancestral globin gene by duplication — early molecular evidence that gene families are built by copying.
• Ohno, Evolution by Gene Duplication (1970). Susumu Ohno synthesized these threads into a theory: selection can only conserve or refine existing functions, so evolutionary novelty demands a redundant copy free to mutate. He also proposed that vertebrates underwent two rounds of whole-genome duplication (the seed of the 2R hypothesis). The book reframed how biologists think about where new genes come from.
• Antifreeze from trypsinogen (Chen, DeVries, Cheng, 1997). Sequencing showed the antifreeze glycoprotein gene of Antarctic notothenioid fish evolved from a pancreatic trypsinogen gene, with the ice-binding repeats amplified from a small region — a clean, dated case of neofunctionalization after duplication, timed to the freezing of the Southern Ocean roughly 5 to 14 million years ago.
• The DDC model (Force, Lynch, Postlethwait et al., 1999). Analysis of duplicated zebrafish genes led to the duplication–degeneration–complementation model, which showed that complementary loss-of-function mutations in regulatory regions can preserve both duplicates without any new adaptive function — putting subfunctionalization on rigorous footing and explaining the surprisingly high retention of teleost duplicates.
• Lenski's long-term E. coli experiment. In the citrate-utilizing (Cit+) lineage that arose after ~31,500 generations, the key innovation involved a tandem duplication that placed a normally silent citrate transporter (citT) under an aerobically active promoter — real-time evolution of a new function enabled by duplication, observed in the lab.