henhao
Sign in
GeneticsbioRxiv
Preprint

Parallel evolution under constraint shapes echinocandin resistance in Candida auris

Nicholas C. CauldroniD, Erika N. DortiD, Garrett Weeks, Phillip David RogersiD, Christina A. CuomoiD

PDFDOICode
Post to X
Threads on XPost to BlueskyThreads on Bluesky
In-depth full-text digest available · Jump to digest ↓

Abstract

Drug resistance emerges repeatedly in outbreaks of Candida fungal pathogens, but little is known about its origins or persistence. Here, we investigated the evolutionary processes shaping echinocandin resistance in Candida auris, a globally emerging and predominantly clonal fungal pathogen. Genome-wide association across over 600 isolates identified mutations in the {beta}-1,3-glucan synthase gene FKS1 as the most significant driver of resistance to an echinocandin drug. Ancestral reconstruction of this population traced shared resistance mutations among small groups typically consisting of 2-3 closely related isolates, but clusters could include up to 16 isolates. Nearly all resistant clusters consisted of isolates collected in the same year and region, consistent with local transmission. To further examine population-level selection, we measured adaptive signatures in FKS1 and the highly diverged paralog FKS2 across 22,000 genomes. This revealed excess nonsynonymous polymorphisms in FKS1, primarily due to independent, recurrent mutations at resistance hotspots, consistent with parallel evolution and incomplete fixation of adaptive alleles. In FKS2, there is no evidence of hotspots and little support for diversifying selection. Together, these results indicate that resistance mutations emerge under strong genetic constraint, with adaptation restricted to only one FKS homolog and predominantly at mutational hotspots.

02 /Specification & Classification

Click any tag to browse matching papers
Citations
0
OpenAlex Topics

In-depth summary

✦ Multi-agent verified full-text synthesis

Claude Sonnet 4.6 · xhigh effort · 8 Nov 2026 · from the full text of version 2 (3 Sept 2026) · CC BY 4.0

Across 681 matched isolates and 22,388 genomes, high-level micafungin resistance in Candida auris arises through recurrent FKS1 hotspot mutations amid strong purifying selection, and FKS2 appears dispensable.

Earlier descriptions of echinocandin resistance in C. auris rarely separated newly arising mutations from inherited ones, and the contribution of the FKS2 paralog was unestablished. This work quantifies independent origins, local transmission, and selection on both genes.

Key results

  1. ShownFKS1 is the main genetic driver of micafungin resistance. A gene-level burden test of rare variants gave very strong FKS1 associations, and ASP1 was the only other significant gene, with an opposite-sign effect.

    GEMMA linear mixed model on 669 single-sample isolates; no single FKS1 variant exceeded allele frequency 0.05, so rare variants were collapsed per gene. ASP1 had a negative beta, i.e. possibly reduced resistance. · Identification of micafungin resistance…; Genome-wide association study

  2. ShownResistance mutations in FKS1 hotspots arise repeatedly and independently, with a subset expanded by local transmission. S639F appears to have emerged most recurrently and is the only variant found in all three major clades.

    58 isolates carry a hotspot mutation (clades I, III, IV); about half of hotspot missense changes are independent origins. One Clade IV S639P cluster has 16 isolates from Maracaibo, Venezuela, 15-47 SNPs apart. · Recurrent mutation and local transmission of…; Figure legends

  3. ShownFKS1 but not FKS2 shows positive selection, concentrated at hotspots, against a background of strong purifying selection. Residues 635 and 639 are the most recurrently mutated sites.

    22,388 isolates: pN/pS 1.4 - 2.7 for FKS1 vs 0 - 0.9 for FKS2; excluding FKS1 hotspots lowers it to 0.7 - 0.9. Minimum recurrent mutations: 52 (residue 635) and 159 (residue 639). · Diversity and selection on FKS genes; FKS gene evolution among Saccharomycete…; Tables

  4. SuggestedFKS2 carries multiple independent putative loss-of-function alleles in natural isolates, and its disruption in a clade I strain left micafungin and caspofungin susceptibility unchanged, so it appears dispensable for echinocandin response.

    11 putative loss-of-function alleles in FKS2 vs none in FKS1 (nine unique frameshift indels in 39 isolates). Two independent FKS2 disruptants showed no MIC change for micafungin (broth microdilution, strips) or caspofungin (strips). · Loss-of-function mutations in FKS2; Antifungal susceptibility testing (AFST)

  5. ShownC. auris FKS2 is an anciently diverged paralog of FKS1 and shares a more recent common ancestor with the non-resistance-linked C. albicans FKS2 than with the resistance-linked C. glabrata FKS2.

    Genetic distance between C. auris FKS1 and FKS2 is 2.1 substitutions per site, versus 0.1 between C. glabrata FKS1 and FKS2 and 1.6 between C. glabrata FKS1 and FKS3. Codon-model tests indicate strong purifying selection on CTG-clade FKS1. · FKS gene evolution among Saccharomycete…

Methods, data and reproducibility

A genome-wide association (GEMMA linear mixed model) on a micafungin MIC-derived susceptibility index, ancestral reconstruction on clade phylogenies, pN/pS and CODEML selection analyses of FKS1/FKS2, and a CRISPR disruption of FKS2 followed by susceptibility testing.

  • Public C. auris short-read genomes matched to micafungin MIC data (one isolate per case) · Candida auris · n = 681 isolates (669 or 623 after filtering)
  • Global C. auris genomes or FKS-aligned reads from NCBI Pathogen Detection (as of September 22, 2025) · Candida auris · n = 22,388 isolates
  • Newly sequenced isolates reported to be micafungin-resistant without FKS1 mutations (B11784, B12779), plus deposited CDC isolates · Candida auris · n = 2 resequenced isolates · PRJNA722434; PRJNA493622; PRJNA638416
  • Clade I strain 1cF used for FKS2 disruption and susceptibility testing · Candida auris · n = 2 independent disruptants; biological triplicate
  • Genomes of echinocandin-resistance-reported Saccharomycotina for FKS homolog phylogeny and selection (FungiDB release 68 or NCBI) · multiple Saccharomycotina yeasts
Methods & tools
GEMMA v0.98.5, GATK v4.5.0.0, BWA-mem v0.7.18, SnpEff v5.2a, FastTree v2.2.0, TreeTime v0.11.4, RECUR v1.1.0, mkado v0.5.0, OrthoFinder v3.0.1b1, CODEML (PAML 4.10.10), RAxML-ng v1.2.2, AlphaFold 3 Server
Code
https://github.com/Neato-Nick/Cauris_FKS_evolhttps://github.com/broadinstitute/broad-fungalgrouphttps://github.com/jsharbrough/protTrim2CDS
Software & environment
Python v3.11 and R v4.5.1; GEMMA v0.98.5, GATK v4.5.0.0, TreeTime v0.11.4, PAML 4.10.10, MAFFT v7.525, RAxML-ng v1.2.2; reference genome B11205.
Reproducibility gaps
No compute or runtime is stated. Per-isolate GWAS inputs and MICs sit in supplementary tables. Some analysis scripts were written with AI coding assistance (Claude Code, Sonnet 4.6) and reviewed by the authors.
Limitations (6)
  • Few isolates carry FKS2 variants with MIC data, so sampling a larger population could reveal rarer FKS2 evolutionary events. · stated by the authors
  • The high SNP counts among clade IV isolates may have been influenced by mapping to a clade I reference genome, which bears on how its transmission clusters are read. · stated by the authors
  • Fitness costs of FKS1 mutations were not measured, so the explanation of reversions through fitness penalties remains inferred; additional FKS1 alleles need direct testing. · stated by the authors
  • Selection estimates for the C. auris FKS2 sub-family and a joint three-family alignment could not be computed because synonymous substitutions were saturated. · stated by the authors
  • The FKS2 dispensability test used one clade I background (strain 1cF) and only micafungin and caspofungin, so other clades, other echinocandins and in vivo conditions remain untested. · noted in this summary
  • MICs were pooled from different studies and assay ranges and converted to a susceptibility index with a tentative breakpoint; residual inter-laboratory variation could affect GWAS power. · noted in this summary

Who should read:Read this if you study antifungal resistance genomics, parallel evolution, or population-genomic selection tests on drug-target paralogs in Candida auris.

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
THISGifford '26Ross '25Sharma '20Carolus '20Kordalewska '23Burrack '22
85% semantic matchMicrobiology
Hugh Gifford, Arnab Pradhan, Ben A Caswall+10 more·bioRxiv·Jan 18, 2026· 1 citations