henhaobio
Sign in

Communications BiologyEvolutionary BiologyPeer-reviewed journal article

Population genomics reveals an ancient origin of heartworms in canids

Rosemonde Isabella PoweriD, Swaid AbdullahiD, Heather S. WaldeniD, Guilherme G. VerocaiiD, Tiana L. SandersiD, Joe L. Luksovsky, Andrew R. MoorheadiD, Michael Timothy Dzimianski, Jeremy M. FosteriD, Michelle L. MichalskiiD, Alicia RojasiD, Samuel Candanedo ChacóniD, Georgiana DeakiD, Andrei Daniel MihalcaiD, Patrizia DanesiiD, Elias PapadopoulosiD, Piyanan TaweethavonsawatiD, Dung Thi BuiiD, Đỗ Ngọc ÁnhiD, Reuben S.K. Sharma, SS HoiD, Stephen R. DoyleiD, Jan R ŠlapetaiD

PDFDOIPreprintCode
In-depth full-text digest available · Jump to digest ↓

Abstract

Heartworms (Dirofilaria immitis) are parasitic nematodes that cause significant cardiopulmonary-associated morbidity and mortality in canids worldwide. The global spread of heartworms is believed to have occurred alongside the dispersal of domesticated dogs. To test this theory, we analysed the genomes of 127 specimens collected from mammalian carnivore hosts across four continents. Here we show distinct genetic differences between heartworms from different continents, indicating a more ancient dispersal in canid hosts than previously recognised. Using admixture analyses, we find an Asian origin for Australian heartworms, consistent with the arrival of dingoes thousands of years ago. Finally, the genetic relatedness between European and Central American heartworms suggests that modern dispersal, likely associated with human colonisation of the Americas by Europeans, occurred with domesticated dogs. This work sheds light on the population dynamics and deep evolutionary history of a globally widespread parasite of veterinary significance.

In-depth summary

Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 2 (10 Jul 2025) · CC BY 4.0

Whole-genome sequencing of 127 adult Dirofilaria immitis from four continents reveals continental population structure predating dog domestication, an Asian source for Australian worms, and recent Europe-Central America spread.

Heartworms are usually thought to have spread worldwide with human-moved dogs over the past few centuries, which would make them one homogeneous population, but earlier genetic studies used few worms from few regions. Broad genome sampling tests this human-centred view and informs surveillance and drug-resistance control that must account for regional differences between parasite populations.

Key results

  1. ShownHeartworm nuclear genomes split into distinct continental populations. The signal is genome-wide, appearing on every chromosome including X, and regional diversity differs sharply. This contradicts the idea of a single homogeneous global population.

    PCA of 196,550 autosomal SNPs from 124 worms: four continent-matched clusters (PC1 12.6%, PC2 10.69%), same pattern per chromosome; Pi differed between all regions (Wilcoxon P < 0.001) except Europe vs Central America; highest USA Pi 7.39×10^-4, lowest Asia 2.89×10^-4; Australia 5.61×10^-4 vs prior estimate 3.9×10^-5. · Nuclear variants reveal distinct continental…; Figure legends

  2. SuggestedEffective population size (Ne) histories diverge between continents before dog domestication. They track ancient climate shifts, and later gene flow between Asian and USA worms points to dispersal with wild canids across Beringia.

    SMC++ over 1 million years (2.5-year generation time, C. elegans rate 2.7×10^-9): Ne rises near the 116–130 kya warm period and falls during 58–72 kya cooling; curves differ >40 kya. TreeMix Asia-USA migration edge (delta m = 3.4); D-statistics show excess USA-Asia allele sharing (|Z| > 3, 21,088 SNPs, D. ursi outgroup). · Coevolution of canids and heartworms…; Population demography; Figure legends

  3. SuggestedAustralian heartworms are genetically closest to Asian worms. This supports introduction from Asia, plausibly with dingoes, rather than with European colonisation. Queensland worms are highly admixed, and Sydney worms form two separate clusters.

    NGSadmix K = 6 shows shared Asian ancestry; TreeMix gives mutual monophyly of Asian and east-coast Australian samples; AUS-ASIA DXY 5.13×10^-4, second lowest pair; D-statistics |Z| > 3; Ne trajectories overlap until 10–20 kya; Australian PCA of 116,492 SNPs separates Sydney into two clusters. · The origin and spread of heartworms in Australia; Nuclear variants reveal distinct continental…; Figure legends

  4. SuggestedEuropean and Central American heartworms share recent ancestry. This fits a dispersal after dog domestication, plausibly with dogs moved during the European colonisation of the Americas.

    EUR-CENAM DXY 4.95×10^-4, the least divergent pair; shared NGSadmix ancestry at K = 2, 3 and 5; Pi not significantly different between these regions; D-statistics show excess Central America-Europe allele sharing (|Z| > 3); SMC++ Ne curves overlap, then diverge around 500 years ago. · Modern heartworm dispersal between Europe and…; Nuclear variants reveal distinct continental…; Figure legends

  5. SuggestedHeartworm genetic variation follows geography, not host species. Worms from cats, foxes and wild carnivores sit among local dog-derived worms. Worm burdens compiled from the literature suggest canids were the ancestral primary hosts.

    Non-dog samples (cats n = 4, foxes n = 4, single ferret, jackal, leopard, wildcat) cluster with same-region dog worms in PCA; ferret worm Pi 1.21×10^-3; generally low between-host FST; literature burdens in wolves and coyotes resemble dogs, while jackals (P ≤ 0.05) and foxes (P ≤ 0.001) carry fewer worms than dogs (Kruskal-Wallis with Dunn's test). · Inference of transmission between diverse…; Genomic diversity of heartworms from multiple…; Figure legends

Methods, data and reproducibility

Adult worms from dogs and other carnivores were whole-genome sequenced (NovaSeq 6000), mapped to a combined D. immitis plus dog reference and jointly genotyped with GATK. Structure, diversity, admixture, gene flow and Ne history were then inferred with PCA, pixy, NGSadmix, TreeMix, D-statistics and SMC++.

  • Newly sequenced adult heartworms from Australia, USA, Panama, Greece, Thailand, Italy, Romania, Costa Rica and Malaysia, plus five technical replicates · Dirofilaria immitis · n = 120 adult worms (Australia 43, USA 37, Panama 14, Greece 12) · PRJNA1104412 (SRA BioProject)
  • Full analysed cohort pooling new and public genomes from dogs, cats, foxes, ferret, golden jackal, leopard and wildcat on four continents · Dirofilaria immitis · n = 127 worms; 124 retained in the nuclear PCA
  • Public whole-genome data from adult heartworms collected in Sydney, Australia · Dirofilaria immitis · n = 5 · SRR13154013-7
  • Public whole-genome data from one adult heartworm each from Georgia (USA) and Pavia (Italy) · Dirofilaria immitis · n = one adult per accession · ERR034940; ERR034941
  • Outgroup filarial worms: D. repens from Italian dogs, Dirofilaria sp. 'Thailand II' from humans in Vietnam, D. ursi from an American black bear · D. repens; Dirofilaria sp. 'Thailand II'; D. ursi · n = 2, 2 and 1 respectively
  • Combined mapping reference: dimmitis_WSI_2.2 (with mitochondrial and Wolbachia genomes) plus domestic dog genome · D. immitis; Canis lupus familiaris · dimmitis_WSI_2.2; GCA_014441545
  • East-coast Australian subset (Sydney to Lockhart River) used for within-country structure, Pi, DXY and FST · Dirofilaria immitis · n = 48
  • Literature-compiled adult worm burdens and microfilaremia reports across carnivoran hosts (Supplementary Table 1) · Carnivora (dogs, wolves, coyotes, jackals, foxes, others)
Methods & tools
NovaSeq 6000 whole-genome sequencing (Novogene), Monarch Genomic DNA Purification Kit, Trimmomatic, FastQC, MultiQC, Nextflow mapping-helminth pipeline with Minimap2, GATK HaplotypeCaller / CombineGVCFs / GenotypeGVCFs, VCFtools, bcftools, PLINK, SNPRelate (PCA), pixy (Pi, DXY, FST), NGSadmix (ANGSD) with CLUMPAK, TreeMix with OptM, AdmixTools qpDstat (D-statistics), SMC++
Software & environment
FastQC 0.11.8, MultiQC 1.17, Trimmomatic 0.39, Nextflow mapping-helminth v1.0.9, Minimap2 2.17, GATK 4.1.4.1, Samtools 1.14, VCFtools 1.16, SNPRelate 1.38.0, pixy 1.2.10.beta2, TreeMix 1.12, AdmixTools 5.1, PLINK 1.90b6.18, SMC++ 1.15.4, OptM 1.8
Reproducibility gaps
Only raw FastQ reads are listed as deposited (BioProject PRJNA1104412, no URL). Filtered VCFs are not stated as available, and sample metadata and literature worm-burden data are in supplementary files.
Limitations (5)
  • SMC++ demography uses the C. elegans mutation rate (2.7×10^-9 per site per generation) as a proxy, and the generation time is uncertain (1, 2.5 or 4 years tested). Absolute Ne and event timings may therefore be over- or underestimated. · stated by the authors
  • Non-dog hosts are sparsely sampled. The jackal, leopard and wildcat worms are single specimens, all from Romania, which confounds host comparisons; host-specific adaptation cannot be excluded without wider geographic sampling. · stated by the authors
  • South America and Africa were not sampled, and Asian worms came only from Thailand and Peninsular Malaysia, not Indonesia or Melanesia. This may explain the offset between the estimated Asia-Australia divergence and dingo arrival. · stated by the authors
  • Several regions rest on few newly collected worms (Thailand n = 6, Malaysia n = 1, Costa Rica n = 1, Italy n = 3, Romania n = 3). Regional Pi, Ne trajectories and D-statistics for Asia and Europe may therefore be sensitive to individual samples. · noted in this summary
  • The links to dingo arrival, Bering Land Bridge exposure and Columbus-era dog transport rest on visual overlap between Ne curves and dated events. No formal split-time or migration-time estimates with uncertainty are reported. · noted in this summary

Read this if you study filarial parasite population genomics, the biogeography of vector-borne pathogens, or heartworm surveillance and the spread of drug resistance between continents.

Citations
8
FWCI
39.62
Field percentile
top 1%

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISGandasegui '23Terai '23Gojobori '21Martino '25Hattendorf '25Ciucani '21
65% semantic matchGenomics
Javier Gandasegui, Rosemonde Isabella Power, Emily Curry+6 more·bioRxiv·Apr 25, 2023· 2 citations