Data and R code for: "Avian Malaria Infections Are Not Associated With Host Genetic Diversity: A Meta-Analysis"

This repository contains the dataset and full analysis workflow associated with the publication “Avian Malaria Infections Are Not Associated With Host Genetic Diversity: A Meta-Analysis” (Podmokła et al. 2026, Molecular Ecology, https://doi.org/10.1111/mec.70374). The study synthesizes published results on the relationship between individual genetic diversity and haemosporidian infections (prevalence and intensity) in birds.

The repository includes:

a processed dataset (CSV) containing all effect sizes and metadata used in the meta‑analysis,

a fully annotated R Markdown script (Rmd) allowing full reproducibility of data filtering, statistical analyses, heterogeneity assessments, publication‑bias tests, and figure generation.

Both files correspond exactly to the materials used in the published version of the study.

Files included

1. S1_File.csv — Processed dataset used in the meta-analysis

This file includes all extracted and curated effect sizes (k = 85) from 35 studies, along with study‑level, host‑level, parasite‑level, and methodological metadata.

Main columns include:

Study information: paper ID, year, species name (Latin + English), migratory status, geographic coordinates, latitude category.

Genetic data:

marker type (functional vs. neutral),

specific marker (MHC class I/II, TLRs, β‑defensins, microsatellites, SNPs),

genotyping method (e.g., 454, Illumina MiSeq, cloning, DGGE, RADseq).

Parasite information: parasite genus (Plasmodium, Haemoproteus, Leucocytozoon), lineage (if available), infection metric (prevalence or intensity), detection method (PCR/qPCR or blood smears).

Effect sizes:

Fisher’s Z‑transformed correlation (ES_fisher_z),

sampling variance (ES_var.z),

standard error (ES_zSE),

effect direction and data source category (text, figure, supplementary materials, raw data, author correspondence).

Flags used for the analysis: use_in_analysis, Reason for exclusion, etc.

The dataset corresponds to the final curated dataset used directly in the models presented in the paper and Supplementary Information.

2. R_code_24.06.2025.Rmd — Reproducible analysis pipeline

This R Markdown script contains the full analytical workflow:

Included steps:

1. Environment and data preparation

1.1 Loading required R packages, including:ape, dmetar, dplyr, ggplot2, meta, metafor, orchaRd, rotl.

1.2 Importing and cleaning the dataset (S1_File.csv):

filtering effect sizes included in the main analysis,

formatting categorical variables (infection method, marker type, MHC genotyping method, migratory behaviour),

creating prevalence and intensity subsets.

2. Global effect sizes and heterogeneity

2.1 Prevalence

2.1.1 Calculation of the global random‑effects model for prevalence.

2.1.2 Heterogeneity assessment, including:

τ² and I² estimates,

automated outlier detection,

influence analyses,

visual diagnostics (Baujat plot, influence plots, I² plot).

2.1.3 Publication bias diagnostics:

contour‑enhanced funnel plots,

Egger’s regression test,

trim‑and‑fill analysis,

publication‑year trend analysis with robust variance.

2.2 Intensity

2.2.1 Calculation of the global random‑effects model for infection intensity.

2.2.2 Heterogeneity and influence diagnostics parallel to the prevalence workflow.

2.2.3 Publication bias diagnostics (funnel plot, Egger’s test, trim‑and‑fill, publication‑year trends).

3. Meta‑regression / Subgroup analysis

3.1 Effect of genetic marker type

3.1.1 Subgroup model and meta‑regression for prevalence.

forest plots by marker category.

3.1.2 Subgroup model and meta‑regression for intensity.

forest plots for marker‑type subsets.

4. Full phylogenetically informed models

4.1 Preparation of phylogenetic data

matching species names to the Open Tree of Life taxonomy,

pruning/constructing the phylogeny,

generating a phylogenetic variance–covariance matrix,

aligning the dataset with the tree tips.

4.2 Full model for prevalence

multilevel meta‑regression with random effects for study, effect size, species identity, and phylogeny,

fixed moderators:genetic marker type, parasite genus, infection‑detection method, migratory behaviour, latitude category,

extraction of model summaries, I², and moderator estimates (orchard plots).

4.3 Full model for intensity

analogous modelling framework, adjusted for data availability (e.g., parasitemia method).

5. MHC‑specific analyses

5.1 Prevalence (MHC subset)

full multilevel models including moderators:MHC class, simplified genotyping method, parasite genus, infection method, migratory status, latitude category.

5.2 Intensity (MHC subset)

same structure as above (with migratory behaviour removed when redundant).

Session information

Full R session details (R version, platform, and package versions) ensuring complete reproducibility.

The script allows the user to reproduce exactly all results, tables, and figures from the manuscript and the Supporting Information.