Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 51 of the 51 references it cites were found in a registry or an index; 39 of the 42 sources its claims cite were checked for support.

OrthoFinder: phylogenetic orthology inference for comparative genomics

Open access
Emms, David M., Kelly, Steven · 2019 · Genome Biology

Published version of a preprint

Abstract

Here, we present a major advance of the OrthoFinder method. This extends OrthoFinder’s high accuracy orthogroup inference to provide phylogenetic inference of orthologs, rooted gene trees, gene duplication events, the rooted species tree, and comparative genomics statistics. Each output is benchmarked on appropriate real or simulated datasets, and where comparable methods exist, OrthoFinder is equivalent to or outperforms these methods. Furthermore, OrthoFinder is the most accurate ortholog inference method on the Quest for Orthologs benchmark test. Finally, OrthoFinder’s comprehensive phylogenetic analysis is achieved with equivalent speed and scalability to the fastest, score-based heuristic methods. OrthoFinder is available at https://github.com/davidemms/OrthoFinder .

Overview

CoChat Summary

TL;DR

OrthoFinder is updated to perform fully phylogenetic orthology inference from protein sequences, inferring orthogroups, rooted gene trees, a rooted species tree, gene duplication events, orthologs, and comparative genomics statistics. It reports the highest ortholog inference accuracy on Quest for Orthologs benchmarks while remaining fast and scalable, with some accuracy–speed trade-offs and rooting limitations.

Key findings

  • On Quest for Orthologs 2011_04, the default fastest OrthoFinder was 3–24% more accurate on SwissTree and 2–30% more accurate on TreeFam-A than any other method; other OrthoFinder configurations were a further 1–3% more accurate.
  • On the Species Tree Discordance Tests, all OrthoFinder versions had equal or higher pseudo-F-score than all other methods; the default version scored 0–45% higher on STDT and 10–59% higher on GSTDT.
  • Default OrthoFinder ran in 192 s on 4 fungal species and 1.8 days on 256 species with 16 threads, making it the second quickest method; SonicParanoid was fastest at 1.2 days on 256 species.
  • The novel hybrid duplication-loss-coalescence algorithm raised duplication F-score from 72% with species-overlap to 80%; DLCpar (full) was most accurate at 91.8% but could not process a 4-species dataset (2259 trees, 12,958 genes) in 120 h, while OrthoFinder analyzed 18,651 trees and 948,449 genes from 128 fungi in 141 s.

Methods

  • Input is one FASTA file per species; orthogroups are inferred with the original OrthoFinder algorithm, unrooted gene trees with DendroBLAST, the species tree with STAG, the species tree root with STRIDE, gene trees are rooted using the species tree, and orthologs/duplications are inferred with a novel hybrid species-overlap plus duplication-loss-coalescence algorithm.
  • Benchmarks used Quest for Orthologs 2011_04 on 66 species across Eukarya, Bacteria, and Archaea; simulated flies (12,000 trees, 12 species, 12,032 duplications), primates (7500 trees, 17 species, 16,066 duplications), and metazoa (2000 trees, 40 species, 4967 duplications); performance was tested on 4–256 fungal proteomes.
  • Analyses compared precision, recall, F-score, pseudo-F-score, Robinson-Foulds agreement, and runtime against methods including Forester, DLCpar (full/search), species-overlap, Notung, and other orthology inference tools.

Limitations

  • Stated: STRIDE may fail to identify the species tree root, or only exclude the root from clades, when gene trees contain few duplications; DLCpar (full) was more accurate but not scalable; the default fastest OrthoFinder is less accurate than other OrthoFinder configurations; the authors cannot test all possible multiple sequence alignment and tree inference combinations.
  • Evident: STDT and GSTDT lack ground-truth orthologs, so pseudo-F-score substitutes for standard precision/recall; benchmarking is limited to specific Quest for Orthologs datasets and simulated data; OrthoFinder is second fastest, not fastest, on the 256-species dataset; no competitor method also provides gene trees or identifies gene duplication events, so the data-richness advantage is not an accuracy comparison.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1186/s13059-019-1832-y

Cites

51 of 51 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig. 1 Pairwise similarity score-based ortholog inference can be misled by variable sequence evolution rates. a Phylogenetic tree of a typical gene family. The correct orthologs of species A-gene 1 are not identified. b Species A-gene 2 is misidentified as the ortholog of the genes from species B and C. Left-hand side: gene trees with branches to scale and the true orthology relationships, which can be determined from the gene tree. Right-hand side: Reciprocal best hits (RBH) based on gene similarity scores that are monotonic with branch length and the orthology relationships inferred from these scores using standard heuristics (orthologs inferred using RBHs and co-orthology identified from within species hits better than closest RBH [8, 16]). FP, false positive; FN, false negative
Fig. 1 Pairwise similarity score-based ortholog inference can be misled by variable sequence evolution rates. a Phylogenetic tree of a typical gene family. The correct orthologs of species A-gene 1 are not identified. b Species A-gene 2 is misidentified as the ortholog of the genes from species B and C. Left-hand side: gene trees with branches to scale and the true orthology relationships, which can be determined from the gene tree. Right-hand side: Reciprocal best hits (RBH) based on gene similarity scores that are monotonic with branch length and the orthology relationships inferred from these scores using standard heuristics (orthologs inferred using RBHs and co-orthology identified from within species hits better than closest RBH [8, 16]). FP, false positive; FN, false negativeCC BY · PubMed Central
Fig. 2 The OrthoFinder workflow. The method used for each step is shown by the arrow. Published algorithms are shown in italics and are followed by an asterisk. A dotted blue line connecting with a solid arrow indicates additional data that are used in order to carry out the transformation indicated by the solid arrow. MSA, multiple sequence alignment-based tree inference; DLC, duplication-loss-coalescence. (a) Orthogroup inference using the original OrthoFinder algorithm (an orthogroup is the set of genes descended from a single gene in the last common ancestor of all the species under consideration). (b) Gene tree inference. (c) Species tree inference. (d) Species tree rooting (e) Gene tree rooting (f) Hybrid overlap + DLC analysis of rooted gene trees to infer orthologs and gene duplication events. (g) Illustration of the ortholog results table for the genes in each input species (four main boxes). The horizontal divisions within these show the orthologs for each individual species pair. (h) Illustration of the gene duplication event table showing the location of the gene duplication events mapped to the species tree, the location in the gene tree, the percent retention of the duplicate genes in the sampled species, and the genes descended from the gene duplication event. (i) Comparative genomics statistics
Fig. 2 The OrthoFinder workflow. The method used for each step is shown by the arrow. Published algorithms are shown in italics and are followed by an asterisk. A dotted blue line connecting with a solid arrow indicates additional data that are used in order to carry out the transformation indicated by the solid arrow. MSA, multiple sequence alignment-based tree inference; DLC, duplication-loss-coalescence. (a) Orthogroup inference using the original OrthoFinder algorithm (an orthogroup is the set of genes descended from a single gene in the last common ancestor of all the species under consideration). (b) Gene tree inference. (c) Species tree inference. (d) Species tree rooting (e) Gene tree rooting (f) Hybrid overlap + DLC analysis of rooted gene trees to infer orthologs and gene duplication events. (g) Illustration of the ortholog results table for the genes in each input species (four main boxes). The horizontal divisions within these show the orthologs for each individual species pair. (h) Illustration of the gene duplication event table showing the location of the gene duplication events mapped to the species tree, the location in the gene tree, the percent retention of the duplicate genes in the sampled species, and the genes descended from the gene duplication event. (i) Comparative genomics statisticsCC BY · PubMed Central
Fig. 3 Summary of OrthoFinder analysis of a set of Chordata species: Ciona intestinalis, Danio rerio, Oryzias latipes, Xenopus tropicalis, Gallus gallus, Monodelphis domestica, Mus musculus, Rattus norvegicus, Pan troglodytes, and Homo sapiens. Bar charts and heat map contain data for each species, aligned to the corresponding species in the tree in a. a The species tree inferred by STAG and rooted by STRIDE. b Percentage of genes from each species assigned to orthogroups. c The number of species-specific orthogroups. d The number of genes with orthologs in any/all species. e Heat map of the number of orthogroups containing each species pair (top right) and orthologs between each species (bottom left). f Ortholog multiplicities for two species, C. intestinalis and H. sapiens, with respect to all other species. g The number of gene duplication events on each terminal branch of the species tree. h The number of duplications on each branch of the species tree and retained in all descendant species. OG, orthogroup; sp., species; spp., species (plural); dups., gene duplication events
Fig. 3 Summary of OrthoFinder analysis of a set of Chordata species: Ciona intestinalis, Danio rerio, Oryzias latipes, Xenopus tropicalis, Gallus gallus, Monodelphis domestica, Mus musculus, Rattus norvegicus, Pan troglodytes, and Homo sapiens. Bar charts and heat map contain data for each species, aligned to the corresponding species in the tree in a. a The species tree inferred by STAG and rooted by STRIDE. b Percentage of genes from each species assigned to orthogroups. c The number of species-specific orthogroups. d The number of genes with orthologs in any/all species. e Heat map of the number of orthogroups containing each species pair (top right) and orthologs between each species (bottom left). f Ortholog multiplicities for two species, C. intestinalis and H. sapiens, with respect to all other species. g The number of gene duplication events on each terminal branch of the species tree. h The number of duplications on each branch of the species tree and retained in all descendant species. OG, orthogroup; sp., species; spp., species (plural); dups., gene duplication eventsCC BY · PubMed Central
Fig. 4 a–l Quest for Orthologs 2011_04 benchmarks (see [1]) on 66 species across Eukarya, Bacteria and Archaea for ortholog inference methods. Dotted line shows Pareto frontier. Data for graphs are in Additional file 1: Table S1. a, b F-score on SwissTree and TreeFam-A tests. c “Pseudo-F-score” across the two Species Tree Discordance Tests (STDT). d “Pseudo-F-score” across the four Generalized Species Tree Discordance Tests (GSTDT). e–f Agreement of orthologs SwissTree/FreeFam-A trees g-h Benchmarks across the STDT & GSTDT. X-axis: Total fraction of randomly selected genes with predicted orthologs in a predefined set of species for the two STDTs & four GSTDTs respectively. Y-axis: Average (1 – normalised Robinson-Foulds distance) between gene tree for putative orthologs and the known species tree across the two STDT & four GSTD respectively. The four individual GSTDTs and two individual STDTs are shown in Additional file: 1 i-l Zoom in of plots e-h. See Methods section “Ortholog Benchmarking” for details of Quest for Orthologs benchmarks. m Runtime for each method with 4-256 input Fungi proteomes. n Results returned by methods, a multi-species orthogroup is the set of genes descended from a single gene in the last common ancestor of three or more species
Fig. 4 a–l Quest for Orthologs 2011_04 benchmarks (see [1]) on 66 species across Eukarya, Bacteria and Archaea for ortholog inference methods. Dotted line shows Pareto frontier. Data for graphs are in Additional file 1: Table S1. a, b F-score on SwissTree and TreeFam-A tests. c “Pseudo-F-score” across the two Species Tree Discordance Tests (STDT). d “Pseudo-F-score” across the four Generalized Species Tree Discordance Tests (GSTDT). e–f Agreement of orthologs SwissTree/FreeFam-A trees g-h Benchmarks across the STDT & GSTDT. X-axis: Total fraction of randomly selected genes with predicted orthologs in a predefined set of species for the two STDTs & four GSTDTs respectively. Y-axis: Average (1 – normalised Robinson-Foulds distance) between gene tree for putative orthologs and the known species tree across the two STDT & four GSTD respectively. The four individual GSTDTs and two individual STDTs are shown in Additional file: 1 i-l Zoom in of plots e-h. See Methods section “Ortholog Benchmarking” for details of Quest for Orthologs benchmarks. m Runtime for each method with 4-256 input Fungi proteomes. n Results returned by methods, a multi-species orthogroup is the set of genes descended from a single gene in the last common ancestor of three or more speciesCC BY · PubMed Central
Fig. 5 a Duplication F-score, on simulated gene trees. b Runtime to analyse all trees from the 4 to 128 species Fungi datasets (see Methods), a maximum time of 120 hours (4.3x10 seconds) was allowed. DLCpar (full) did not complete the smallest dataset in this time limit and so only the lower bound for the first time point is shown. c-d Precision and recall
Fig. 5 a Duplication F-score, on simulated gene trees. b Runtime to analyse all trees from the 4 to 128 species Fungi datasets (see Methods), a maximum time of 120 hours (4.3x10 seconds) was allowed. DLCpar (full) did not complete the smallest dataset in this time limit and so only the lower bound for the first time point is shown. c-d Precision and recallCC BY · PubMed Central

Verification

Claims the paper makes

23 Valid · 10 Not valid · 1 Not checked · 62 Own results and interpretations

39 more claims · 8 with their cited sources checked

Create a free account to see all 96 claims and every source verdict.

See all claims — free account
  • Probably not a substantial claim. 18
  • It may state two separate claims as one. 40
  • The quote only partly states this claim. 2

Create a free account to see these claims.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

4 more claims about this paper

Create a free account to see every claim other papers make about this one.

See all claims — free account

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

2 of 2

Citations

Cites

Show all 49 references

Cited by

See all 7,958

Metadata

Authors
Emms, D. M., & Kelly, S.
Affiliations
University of Oxford (GB)
Corresponding author
Kelly, Steven
Venue
Genome Biology
Published
2019-11-14
Volume
20
Issue
1
Pages
238-238
ISSN
  • 1465-6906 (ISSN-L)
  • 1474-760X
OpenAlex
W2988396470
Cited by
7,676
References
49
Citations by year
2026: 1354 · 2025: 1592 · 2024: 1601 · 2023: 1155 · 2022: 984 · 2021: 689 · 2020: 263 · 2019: 32
FWCI
168.65
Type
journal article
Language
en
Open access
Licence
Locations
  • Genome biology (publishedVersion, cc-by)
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
  • ERC (637765)
  • Royal Society (UF140484)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Genomics and Phylogenetic Studies
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Genetics › Chromosomal and Genetic Variations
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Bioinformatics and Genomic Networks
Keywords
OrthoFinder, orthogroups, phylogenetic inference, gene trees, gene duplication, comparative genomics, benchmarking
MeSH terms
Algorithms, Animals, Phylogeny, Software, Genomics

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-11-02, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-11-02 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: OrthoFinder: phylogenetic orthology inference for comparative genomics. https://app.cochat.ai/papers/10.1186/s13059-019-1832-y
The paper
Emms, D. M., & Kelly, S. (2019). OrthoFinder: phylogenetic orthology inference for comparative genomics. Genome Biology. https://doi.org/10.1186/s13059-019-1832-y

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.