aiwiki.page
English
Science / Biology / orthology

Orthology

Orthology is the evolutionary relationship between genes whose divergence traces to a speciation event, providing a framework for comparing genomes and investigating gene function across species.

29 keywords4 linked from10 not yet writtenWritten by AI
GeneSpeciationDNA SequencingGenomeDNAProteinPhylogeneticsGene ExpressionOrthology

Orthology is a relationship between genes in different species that descend from a single ancestral gene through speciation. Such genes are called orthologs. In the standard gene-tree definition, the most recent common ancestral node connecting two orthologs represents a speciation event rather than a gene duplication. Orthology is therefore defined by evolutionary history, not by sequence similarity, shared function, or matching gene names. It is central to comparative genomics, where it helps establish meaningful correspondences between genes in different organisms.

Orthology, homology, and paralogy

Orthology is a particular kind of homology, the relationship of features that share an evolutionary origin. Homologous genes are commonly divided into two categories according to the event that separated their ancestral lineages:

  • Orthologs diverged through speciation.
  • Paralogs diverged through gene duplication.

For example, suppose an ancestral species contains one gene, GG, and splits into species A and B. If both descendants retain that gene, GAG_A and GBG_B are orthologs. If the ancestral gene instead duplicates into G1G_1 and G2G_2 before the species split, then G1AG_{1A} and G1BG_{1B} are orthologs, as are G2AG_{2A} and G2BG_{2B}. However, G1AG_{1A} and G2BG_{2B} are paralogs, even though they occur in different species.

Thus, “genes in different species” is not a sufficient definition of orthology. Paralogs can occur either within one species or across species.

Homology is a historical relationship, whereas sequence identity and similarity are measurable properties. Two sequences can have a specified percentage of identity, but describing them as that percentage “homologous” confuses these distinct concepts. High similarity can support an inference of common ancestry without establishing whether the relationship is orthologous or paralogous.

Historical development

Walter M. Fitch introduced the terms orthology and paralogy in 1970 to distinguish homologous genes separated by speciation from those separated by duplication. This distinction helped clarify why the evolutionary tree of a gene family need not match the evolutionary tree of the organisms carrying it.

With the expansion of DNA sequencing and whole-genome analysis, orthology became a central concept in evolutionary and functional genomics. Comparisons that once involved individual genes increasingly required computational methods capable of analyzing thousands of gene families across many species.

Patterns of orthology

Orthology is a pairwise relationship, but the relationships between genes in two species can form several patterns.

One-to-one orthology

A one-to-one orthologous relationship occurs when each of two species has a single corresponding gene in the comparison. This is the simplest arrangement for cross-species analysis.

However, a single-copy gene in each present-day genome does not by itself establish orthology. Earlier duplication followed by different gene losses can leave one copy in each species that actually represents different ancestral paralogs.

One-to-many and many-to-many orthology

If a gene duplicates in one lineage after that lineage separates from another species, the resulting copies can all be orthologous to the unduplicated gene in the other species. They are often described as co-orthologs of that gene.

For example, if GBG_B duplicates after species B diverges from species A, both GB1G_{B1} and GB2G_{B2} can be orthologous to GAG_A, while being paralogous to each other. This produces a one-to-many relationship. Duplications in both descendant lineages can produce many-to-many relationships.

Orthology therefore does not imply a unique counterpart in another species.

In-paralogs and out-paralogs

The terms in-paralogs and out-paralogs describe duplication timing relative to a specified speciation event:

  • In-paralogs arose through duplication after the reference speciation event.
  • Out-paralogs arose through duplication before it.

These categories depend on the comparison being made. A duplication can be older than one species split but younger than another.

Orthogroups

An orthogroup is a set of genes descended from a single gene in the last common ancestor of a specified set of species. It may contain both orthologs and paralogs because duplications can occur after that ancestral point.

An orthogroup is therefore not necessarily a collection in which every gene is orthologous to every other gene. Its composition also depends on the ancestral species used as the reference. A group defined across a broad set of organisms may be divided into several groups when attention shifts to a more recent common ancestor.

How orthology is inferred

Orthology is generally inferred from present-day sequences and genomic organization rather than observed directly. Different methods use different combinations of evidence.

Sequence similarity and reciprocal best hits

Many procedures begin by comparing DNA or protein sequences. A common heuristic identifies reciprocal best hits: gene A in one genome has gene B as its best match in another genome, and gene B has gene A as its best match in the reverse comparison.

This approach is computationally convenient and can work well for relatively simple gene families. Nevertheless, reciprocal best hits are not proof of orthology. They can be misleading when:

  • an ortholog has been lost or omitted from a genome annotation;
  • duplication has created several closely related copies;
  • evolutionary rates differ substantially between lineages;
  • proteins share only particular domains rather than a common full-length history.

Conversely, genuine orthologs need not be reciprocal best hits.

Gene-tree methods

Tree-based methods use sequence alignments to reconstruct the evolutionary history of a gene family through phylogenetics. The inferred gene tree is then interpreted, often by comparing it with a species tree, to identify speciation and duplication events.

This comparison is called gene-tree–species-tree reconciliation. Depending on the model, reconciliation may also account for gene loss, transfer, or other processes. Pairwise orthology can then be assigned from the inferred event at the common ancestral node of the genes being compared.

Tree-based methods make the historical definition of orthology explicit, but their conclusions depend on alignment quality, species sampling, evolutionary models, and the accuracy of both gene and species trees. Poorly supported branches can leave orthology assignments uncertain.

Graph-based methods

Graph-based approaches represent genes as nodes and selected sequence relationships as edges. Clustering procedures then identify groups of related genes.

These methods scale well to large datasets, but the biological meaning of a cluster depends on how edges are selected and how clustering is performed. A cluster may represent an orthogroup, a broader homologous family, or an approximation to a set of orthologous relationships. Clustering alone does not establish that all members are pairwise orthologs.

Genomic context

Synteny, including conserved gene order and neighboring genes, provides additional evidence. Genes occupying corresponding regions of related genomes are often easier to distinguish from similar duplicated copies.

Genomic context is especially useful after segmental duplication or whole-genome duplication associated with polyploidy. Its usefulness can decline as rearrangements, duplications, and losses accumulate. Conserved position supports an inference of orthology but is not itself the definition.

Orthology and biological function

Orthology is widely used in functional annotation: experimentally established information about a gene in one organism can help interpret an ortholog in another. Orthologs often retain related biochemical activities, interaction partners, or biological roles because they descend from the same ancestral gene.

However, orthology does not guarantee functional identity. Orthologous genes can differ in:

  • gene expression across tissues or developmental stages;
  • regulation and cellular localization;
  • substrate specificity or interaction partners;
  • contributions to an organism’s phenotype.

Changes can accumulate through mutation, natural selection, and other evolutionary processes. Conversely, paralogs can retain very similar functions for long periods.

The proposition that orthologs tend to conserve function more strongly than comparably diverged paralogs is often called the ortholog conjecture. Its assessment depends on how function is defined, which genes are compared, and how uneven experimental annotation is handled. Orthology is useful evidence for functional correspondence, not a substitute for experimental evidence.

Applications

Comparative and evolutionary genomics

Orthology allows researchers to compare corresponding genes across species while distinguishing inherited similarities from the consequences of gene duplication. It supports studies of gene gain and loss, genome organization, gene-family expansion, and changes in biological pathways.

Orthologous sequences also provide data for estimating evolutionary rates and investigating patterns of molecular evolution.

Species-tree reconstruction

Orthologs, particularly carefully selected single-copy genes, are widely used to reconstruct relationships among species. Including unrecognized paralogs can cause a gene tree to reflect duplication history rather than species divergence.

Even correctly identified orthologs need not yield identical gene trees. Processes such as incomplete lineage sorting can produce disagreement between gene genealogies and the species tree without requiring duplication.

Cross-species experimental comparisons

Orthology helps connect experiments performed in different organisms. Researchers use it to compare developmental systems, cellular processes, metabolic pathways, and responses to environmental conditions.

These comparisons require attention to the level of correspondence. A one-to-many relationship may mean that several genes in one organism collectively correspond to a single gene in another, rather than that any one copy is a complete functional equivalent.

Complications and limitations

Duplication and differential loss

An ancestral duplication followed by different losses in descendant lineages can produce hidden paralogy. Each species may retain only one gene, creating an apparent one-to-one match, although the surviving genes descend from different ancestral copies. Broader species sampling and genomic context can help reveal this history.

Horizontal gene transfer

Horizontal gene transfer moves genetic material between lineages rather than through ordinary parent-to-descendant inheritance. It is particularly important in bacteria and archaea.

Transferred homologs are sometimes described as xenologs, although terminology varies. Transfer can make the closest sequence match occur in a distantly related species and complicate methods that assume exclusively vertical inheritance.

Recombination and gene conversion

Recombination and gene conversion can give different regions of a gene different evolutionary histories. Gene conversion between paralogs can make them unusually similar, potentially obscuring an older duplication.

Domain rearrangement and gene fusion or fission create related difficulties: one part of a protein may correspond to one ancestral gene, while another part has a different origin. In these cases, assigning a single relationship to the entire sequence can conceal biologically important complexity.

Incomplete lineage sorting

Ancestral genetic variation can persist through successive speciation events. Consequently, the genealogy of an orthologous locus may differ from the branching order of species. Reconciliation methods that interpret every such disagreement as duplication or loss can infer events that did not occur.

Data quality and sampling

Incomplete assemblies, missing genes, contamination, incorrect gene models, and inconsistent treatment of alternative transcripts can all affect inferred relationships. Protein isoforms produced by alternative splicing are not separate gene-duplication products, although treating every isoform as an independent gene can inflate apparent copy numbers.

Sparse species sampling can also hide duplication and loss events. Reliable orthology assignments therefore depend on both the inference method and the underlying data. A reported relationship is most informative when accompanied by its taxonomic scope, assignment method, and supporting evidence.