Orthologs are versions of the same gene found in different species, inherited from a single gene in the ancestor those species share.1Annual Review of Genetics. Orthologs, Paralogs, and Evolutionary Genomics The gene that helps build a human eye and its counterpart in a mouse can be traced back to one ancestral gene. Telling these inherited counterparts apart from other related genes is what lets scientists carry lessons from a lab animal over to people.
Orthologs Versus Paralogs
Geneticist Walter Fitch drew a line between two kinds of gene relatives based on how they came to exist.2PubMed Central. Functional and evolutionary implications of gene orthology Orthologs come from a species splitting in two. Paralogs come from a gene being copied within one genome.3PubMed Central. Functional and evolutionary implications of gene orthology The names hint at this: “ortho” means exact, and “para” means beside.4PubMed Central. Functional and evolutionary implications of gene orthology So humans carry many paralogs side by side in their own DNA, while their orthologs sit in other species.
A family tree makes the difference easier to hold. Cousins on different branches descend from the same grandparents, much as orthologs descend from one shared ancestor gene. Twins born into one household resemble paralogs made by copying. Both are relatives, but they became related in different ways. Biologists use the word homologs for both groups together, meaning genes that share ancestry.
Hemoglobin shows how this plays out. Humans carry several globin genes, including the alpha and beta chains of hemoglobin, that arose by copying within an ancestral genome, so those genes are paralogs of each other. The human beta-globin gene and the mouse beta-globin gene are a different kind of pair. They descend from the same gene in a common mammal ancestor, so they are orthologs. One gene can have both kinds of relatives at once: orthologs in other species and paralogs in its own genome.
This is why similar sequence does not settle the question. Two genes can look alike because of duplication, not because a species split, and similarity alone cannot tell you which kind of relative you have. The history of how the gene arose decides the label.
Scientists also use finer labels. Copies made before two species split are called outparalogs, and copies made afterward in just one lineage are inparalogs.5Annual Review of Genetics. Orthologs, Paralogs, and Evolutionary Genomics These labels only make sense relative to a particular split between species.6Annual Review of Genetics. Orthologs, Paralogs, and Evolutionary Genomics Genes can also jump between species, especially among microbes, producing look-alikes that can be mistaken for orthologs.7Annual Review of Genetics. Orthologs, Paralogs, and Evolutionary Genomics
The same split can produce different patterns. If a gene has not been copied since two species diverged, each species has one version and the orthologs form a one-to-one pair. If one species copied the gene afterward, a single gene in one species matches two or more in the other. If both species made copies, the relationship becomes many-to-many, and every gene in one group is an ortholog of every gene in the other. That is why “the ortholog” of a human gene is sometimes several genes, and why the answer depends on which two species you compare.
What Happens to Genes After They Are Copied
The most common fate of a gene copy is that one copy gets deleted, returning to a single gene.8PubMed Central. The multiple fates of gene duplications: Deletion, hypofunctionalization, subfunctionalization, neofunctionalization, dosage balance constraints, and neutral variation A new copy usually adds nothing the organism needs, because the original already does the job. Random mutations that damage one copy cause no harm while the other still works, so the damaged copy can slowly become a nonfunctional leftover. Over many generations it is deleted or decays beyond recognition. The surviving pairs we see in genomes are the minority that found some way to be useful.
Sometimes both copies are kept, each working at a lower level so that together they supply enough.9PubMed Central. The multiple fates of gene duplications: Deletion, hypofunctionalization, subfunctionalization, neofunctionalization, dosage balance constraints, and neutral variation Some genes are needed in a particular amount, and two copies at full strength might make too much protein. Once each copy has slipped to a lower level, losing either one would leave the organism short, so natural selection keeps both. A gene pair can persist this way even if neither copy has gained any new ability.
The two copies can also divide the original job between them, or one can evolve a new job.10PubMed Central. The multiple fates of gene duplications: Deletion, hypofunctionalization, subfunctionalization, neofunctionalization, dosage balance constraints, and neutral variation Picture an ancestral gene that worked in both the eye and the gut. After copying, one version may keep the eye role and the other the gut role. Now both are needed, so neither gets lost. Modeling suggests that splitting the job is often a stepping stone toward one copy gaining a new function.11PubMed Central. Subfunctionalization of duplicated genes as a transition state to neofunctionalization
A spare copy is free to change because mutations in one copy are less likely to harm the organism when a backup exists. That freedom lets one copy experiment with a new role, and it is a major way genomes gain new abilities.
Orthologs have less room to wander. Across vertebrates, orthologs separated by a species split changed in more similar ways than paralogs separated by duplication.12PubMed Central. Patterns of gene evolution following duplications and speciations in vertebrates Each ortholog stays the only copy doing its job, so there is less freedom to drift.
Do Orthologs Do the Same Job
Scientists long assumed orthologs share function more than paralogs do, an idea called the ortholog conjecture. It sounds sensible. A gene in a mouse and a human has had a single job to maintain through a long shared history, while paralogs have had the chance to specialize and so should have wandered further. But a sensible idea still needs data, so researchers tested it by comparing how alike the known functions of ortholog pairs and paralog pairs are.
The results were mixed. One large comparison found orthologs, especially one-to-one pairs, are more alike in recorded functions than paralogs, but only modestly.13PubMed Central. Resolving the Ortholog Conjecture: Orthologs Tend to Be Weakly, but Significantly, More Similar in Function than Paralogs A human-mouse comparison found that paralogs within one species predicted function better than orthologs in the other.14PubMed Central. Testing the Ortholog Conjecture with Comparative Functional Genomic Data from Mammals Later work using gene activity measurements found orthologs were generally more alike than paralogs.15PubMed Central. The Ortholog Conjecture Is Untestable by the Current Gene Ontology but Is Supported by RNA Sequencing Data
The disagreement comes largely from the records. Descriptions of what genes do come from published experiments, and some genes and species have been studied far more than others. That uneven attention can make some pairs look more alike simply because the same research groups described them in the same way. Paralogs within one species are often studied by the same laboratory, which can make their records look more similar than they truly are. Measuring activity directly in tissues avoids many of these biases, which helps explain why those results were more consistent.
Activity data show a good deal of conservation. About 84% of human-mouse ortholog pairs are switched on in more similar patterns than random gene pairs.16PubMed. Evolutionary conservation of expression profiles between human and mouse orthologous genes Tissue-specific activity stays similar even between humans and pufferfish, separated about 450 million years.17PubMed Central. Conservation of core gene expression in vertebrate tissues Brain genes are the most consistent across species, while kidney genes differ the most.18PubMed Central. Conservation of core gene expression in vertebrate tissues Whether a gene is limited to certain tissues is still partly shared between humans and flies.19PubMed Central. Tissue-Specificity of Gene Expression Diverges Slowly between Orthologs, and Rapidly between Paralogs
There is a caveat. The way one gene’s message is cut into variants changes quickly: within about 6 million years, these patterns group by species more than by organ.20Science. The evolutionary landscape of alternative splicing in vertebrate species So matching orthologs can still produce different protein versions in different species.
For a reader, that means an ortholog is a strong clue about function, not a guarantee. Experiments are still needed to confirm a shared role.
How Scientists Find Orthologs
One approach links genes that are each other’s closest match between two genomes. Another builds gene family trees and fits them onto the species tree.21PubMed. Inferring orthology and paralogy
The best-match method compares every gene in one genome to every gene in another and looks for pairs that choose each other. Such mutual best matches are a reasonable sign of orthologs because the two genes are more alike than any rival copy. It is quick and easy to run on whole genomes, which makes it widely used. It struggles when a gene has been copied, because the closest match may be a paralog and not the true ortholog.
The tree-based method first groups related genes into a family and draws a tree showing how they branch from one another. Researchers then compare that gene tree to the known species tree to see which branches reflect species splits and which reflect copying. Branches matching a split mark orthologs, and branches matching a copying event mark paralogs. Because it depends on an accurate tree, errors in the tree can lead to missed orthologs.
The two approaches trade off differently. Tree-based methods rarely mislabel a gene, 1 to 7% of the time, but miss 24 to 64% of true orthologs.22PubMed Central. Assessing Performance of Orthology Detection Strategies Applied to Eukaryotic Genomes Simple similarity searches catch nearly all orthologs while wrongly including 50 to 56% that aren’t.23PubMed Central. Assessing Performance of Orthology Detection Strategies Applied to Eukaryotic Genomes Strict methods give trustworthy but incomplete lists, loose methods catch everything but add false matches, and researchers choose based on whether a miss or a mistake costs more. Two widely used tools, Inparanoid and OrthoMCL, kept both error types under 20%.24PubMed Central. Assessing Performance of Orthology Detection Strategies Applied to Eukaryotic Genomes
No database is fully right. Checked against carefully hand-curated gene families, major databases fully recovered only 3 to 22% of groups.25PubMed Central. Orthology prediction methods: A quality assessment using curated protein families None of seven compared databases gave a fully correct classification.26PubMed Central. Conceptual framework and pilot study to benchmark phylogenomic databases based on reference gene trees Faster search programs also miss more matches as species become more distant.27PubMed Central. Quickly Finding Orthologs as Reciprocal Best Hits with BLAT, LAST, and UBLAST: How Much Do We Miss? So it helps to compare results from more than one database, treat one-to-one matches as most reliable, and be cautious with distant species.
Gene neighborhood adds extra evidence. Genes that sit in the same order on chromosomes in two species are likely to be true counterparts.28PubMed. Using shared genomic synteny and shared protein functions to enhance the identification of orthologous gene pairs Using it found 19,357 human-mouse ortholog pairs, 20% more than usual methods.29PubMed. Using shared genomic synteny and shared protein functions to enhance the identification of orthologous gene pairs Neighborhood checks also remove paralogs that slipped into ortholog groups.30PubMed Central. OrthoRefine: automated enhancement of prior ortholog identification via synteny
Why Orthologs Matter for Medicine
Scientists use lab species such as mice, flies, and yeast to study human disease genes through their orthologs.31PubMed Central. An integrative approach to ortholog prediction for disease-focused and other functional studies Mouse data contributes most to finding human disease genes, while other species add little in that task.32PubMed Central. Contribution of model organism phenotypes to the computational identification of human disease genes Mice are mammals with closely matching genes and organs, which fits the closer functional match described earlier.
The workflow is straightforward. A researcher first finds the mouse ortholog of the human gene linked to a disease, using an ortholog database. They then change or switch off that gene in mice and look for signs resembling the human condition. If the mice show similar problems, the mouse can be used to study how the disease develops and to test treatments. The whole approach depends on the ortholog being the right match, which is why correct identification matters.
Sometimes the best model is surprising. Shared ortholog patterns revealed a worm model for breast cancer and a plant model for a hearing and pigment disorder called Waardenburg syndrome.33PubMed Central. Systematic discovery of nonobvious human disease models through orthologous phenotypes A worm or a plant can be a cheaper and faster way to study a gene’s biology than a mammal, and these choices come from looking at shared genes, not at how similar the animals look.
That works because most diseases involve a network of genes working together, not a single gene acting alone. Some of these networks are ancient and carry out basic tasks such as controlling cell growth or building pigment cells. A worm or plant has no breasts or ears, but it can still have the same network doing a related basic job inside its cells. The visible symptoms differ, but the broken genes underneath can match.
Orthologs also help in the search for drugs against germs. Genes that are essential in a germ and present across many pathogens point to broad-spectrum drug targets, and targets are ranked by whether humans lack a matching gene.34PubMed Central. From genetic footprinting to antimicrobial drug targets: examples in cofactor biosynthetic pathways Ten genes were found in all studied disease-causing fungi but not in humans.35PubMed Central. Comparative genomics allowed the identification of drug targets against human fungal pathogens Drugs hitting such genes could spare human cells.
Orthologs Beyond Medicine
Orthologs present as a single copy in each species are often used to build evolutionary trees.36Ecological Research. CUSCO: A Tool for Curating Single‐Copy Orthologs and Extracting Marker Genes for Phylogenetic Tree Construction With Extra Samples When every species has just one copy, the genes line up one to one, which removes the risk of comparing a gene to a paralog. Differences between those copies reflect the time since the species split, so they can be used to work out relationships. Combining many such genes gives a more reliable tree than a few chosen genes, and it reduces the chance that one unusual gene misleads the result.37PubMed Central. Phylogenetic analysis based on single-copy orthologous proteins in highly variable chloroplast genomes of Corydalis
Orthologs also help label new genomes. When a species is newly sequenced, nobody knows what most of its genes do. Gene function labels from studied species can be transferred to orthologs in the new one,38PubMed. Annotated genes and nonannotated genomes: cross-species use of Gene Ontology in ecology and evolution research giving a first draft of what the genome contains that then guides later experiments. The labels are only as reliable as the ortholog match behind them. Tools now assign orthologs quickly to label large gene sets, replacing less precise similarity matching.39bioRxiv. Fast genome-wide functional annotation through orthology assignment by eggNOG-mapper
Some cases are hard. All vertebrates, humans included, descend from an ancestor that doubled its whole genome twice.40bioRxiv. OHNOLOGS v2: A comprehensive resource for the genes retained from whole genome duplication in vertebrates Doubling creates many paralogs at once, and researchers studying fish must decide which copy matches a human gene. Most fish doubled a third time, so one human gene may match several fish genes.41bioRxiv. OHNOLOGS v2: A comprehensive resource for the genes retained from whole genome duplication in vertebrates Genes that don’t make proteins change so fast that matching them across species is much harder.42PubMed Central. Detection of long non–coding RNA homology, a comparative study on alignment and alignment–free metrics