Skip to main content
NCBI home page
Genome Biology and Evolution logoLink to Genome Biology and Evolution
. 2026 Jun 13;18(6):evag139. doi: 10.1093/gbe/evag139

Silent Erosion: Impact of Forest Logging on Genetic Diversity in Tropical Understorey Birds

Simone Messina 1,✉,✉, David P Edwards 2,3, Suzanne Tomassi 4, Suzan Benedick 5, Yong Chee Keita Sin 6, Frank E Rheindt 7,✉,✉, David Costantini 8
Editor: Diego Cortez
PMCID: PMC13312955  PMID: 42287228

Abstract

Anthropogenic habitat disturbance can erode genetic diversity long before population declines become apparent, potentially contributing to extinction debt. However, empirical assessments of genetic erosion following anthropogenic habitat degradation remain scarce. Here, we investigated whether selective logging, the most pervasive anthropogenic disturbance in tropical forests, affects genetic diversity in tropical forest biota. We resequenced whole genomes of 99 understorey songbirds from ten forest specialist species captured in both primary and selectively logged forests of Borneo. The study species are all sedentary and weak flyers but differ in their susceptibility to forest logging. The within-species comparisons of whole genomes between primary and selectively logged forests showed reduced nucleotide diversity and higher levels of inbreeding in disturbed versus undisturbed habitat for a common understorey rainforest inhabitant, Malacocincla sepiaria. We also found a substantial amount of homozygosity in the local population of that same species as well as in another, Malacopteron cinereum, regardless of forest type, indicating reduced standing genetic diversity due to historical bottlenecks. Nine out of ten species, including those considered more sensitive to the environmental conditions of selectively logged forests, showed no reduction in genetic diversity, indicating that selectively logged forests retain substantial conservation value for many rainforest birds. Nonetheless, our findings indicate that genetic erosion can occur within a few decades of habitat disturbance even in apparently common tropical bird species, raising concerns about silent population-genetic processes that might threaten the long-term persistence of seemingly abundant species in selectively logged forests.

Keywords: genetic erosion, genetic diversity, forest logging, population genomics, bird

Graphical Abstract

Graphical Abstract.

For image description, please refer to the figure legend and surrounding text.

Open in a new tab

Significance.

Selective logging is the most pervasive anthropogenic disturbance in tropical forests, but whether it leads to the erosion of genetic diversity in forest organisms is a key unresolved question. By comparing resequenced whole genomes of ten species of understorey songbirds between old-growth primary forests and selectively logged forests in Borneo, we show the first evidence of reduced nucleotide diversity and higher levels of inbreeding in a common rainforest songbird species in selectively logged forests. Our findings raise concerns about silent population-genetic processes that might threaten the long-term persistence of seemingly abundant species in selectively logged forests.

Introduction

Human activity is responsible for a global biodiversity crisis that has resulted in the Sixth Extinction Event (Ceballos and Ehrlich 2023). A realistic evaluation of its impacts is paramount for guiding conservation action and to assess the potential effects this crisis may have on basic planetary processes. However, most efforts to quantify the magnitude of biodiversity loss have focused on direct impacts resulting in visually observable data, measured via such approaches as population censuses or extinction surveys (Edwards et al. 2011; Paez et al. 2022). At the same time, it is known that the impacts of habitat loss or degradation can take many years to crystallize in the form of observable extinctions or declines. They can build up in the background, accruing extinction debt, which only materializes years or decades later through the sudden disappearance of species (Kuussaari et al. 2009).

The mechanisms underlying extinction debt remain poorly understood, but it has been suggested that the resultant time lag in declines or extinctions is intertwined with population-genetic processes (Figueiredo et al. 2019). Reduced survival and reproductive success due to anthropogenic habitat disturbance cause the loss of genetic variants. Such loss of allelic diversity can decrease a species’ adaptive potential, ultimately leading to populations’ inability to deal with rapid environmental change (Gargiulo et al. 2024). In its worst form, diversity loss leads to inbreeding depression, with concomitant repercussions through the accumulation of genetic load (Dussex et al. 2023). These insights suggest that genetic diversity scans may be an invaluable tool of the future to flag populations at risk (DeWoody et al. 2021; Schmidt et al. 2023). And yet, there is a lack of comprehensive studies that have investigated the impact of habitat loss or degradation directly on genetic diversity.

The most pervasive form of anthropogenic disturbance in tropical forest is selective logging. Over 400 million hectares (∼25%) of the remaining tropical forests are zoned for selective harvesting of large, commercially valuable trees (Blaser et al. 2011; FAO 2025). Compared to old-growth undisturbed forests, selectively logged forests are dominated by fast-growing pioneer species, saplings, and elevated abundance of climbing lianas and bamboos (Imai et al. 2019). Although selectively logged forests retain high levels of species richness (Edwards et al. 2011; Putz et al. 2012), the local abundances of species change as interior specialists decline and edge-tolerant species increase in abundance (Costantini et al. 2016). Studies on tropical understorey birds of selectively logged forests found differences in movement behavior, body size, and physiology compared to conspecifics in primary undisturbed forests (Messina et al. 2020a, 2021; Cosset et al. 2021).

In this study, we tested the hypothesis that selective logging leads to loss of genetic diversity in those species that are sensitive to this kind of forest disturbance. We assessed whether apparently nonsensitive species may additionally be subject to the silent effects of genetic diversity loss. We included in our study ten species of understorey songbirds showing increasing, decreasing, or stable trends of relative population abundance (RPA) from old-growth primary forest toward selectively logged forest in Malaysian Borneo (Table 1). The study species belong to the families Pellorneidae (six babbler species), Timaliidae (three babbler species), and Pycnonotidae (one bulbul species). Babblers are strictly insectivorous, while the sole bulbul species included in our study is frugivorous. All our study species are forest specialists, highly sedentary and do not exhibit a strong dispersal capability.

Table 1.

Species names follow the classification by Eaton et al. (2021)

Species scientific name Species common name Sample size Sample size UNL Sample size LOG RPA index Mean genomic coverage (×) Total genomic sites n SNPs
Kenopia striata* Striped wren babbler 10 5 5 −0.26 14.55 1,192,592 377,512
Malacocincla sepiaria Horsfield's babbler 10 5 5 0.13 12.77 652,481 377,512
Malacopteron cinereum* Scaly crowned babbler 10 5 5 −0.15 13.23 804,711 315,455
Pellorneum bicolor Ferruginous babbler 10 5 5 0.15 14.12 996,629 305,600
Pellorneum capistratoides Bornean black-capped babbler 10 5 5 0.24 15.29 881,358 316,355
Pellorneum poliogene Leaflitter babbler 10 5 5 0.13 15.81 394,143 149,932
Cyanoderma bicolor Bicolored babbler 10 5 5 0.16 13.47 956,482 259,208
Stachyris maculata Chestnut-rumped babbler 6 3 3 0.03 13.45 2,508,264 546,464
Stachyris poliocephala Gray-headed babbler 10 5 5 0.23 14.44 1,047,259 280,131
Tricholestes criniger Hairy-backed bulbul 13 7 6 −0.01 14.92 175,572 119,737
Open in a new tab

UNL, primary (unlogged); LOG, logged; n SNPs, number of single nucleotide polymorphisms.

The relative population abundance (RPA) index is negative for species with lower abundance in logged as compared to primary forest and positive for species with higher abundance in logged forest (ranging from −1 to 1). The two species for which we expected a loss of genetic diversity are indicated by asterisks. For all species except T. criniger, SNPs were retained if they were genotyped in at least 90% of individuals; for T. criniger the threshold was 70%.

Results

An average of 298,339 single nucleotide polymorphisms (SNPs) per species were retained after filtering. The species with the highest number of SNPs was Stachyris maculata (n SNPs = 546,464) and with the lowest number Tricholestes criniger (n SNPs = 119,737; Table 1).

Visualizations of principal component analyses (PCA) and admixture analyses showed no clear pattern of segregation of populations between logged and primary forests across all species (Figs. S1 and S2). Levels of nucleotide diversity (π) ranged from 0.009 for Stachyris poliocephala in both primary and logged forests to 0.02 for Malacocincla sepiaria in primary forest (Table S1). The analysis of effect sizes (ES) pointed to a significantly reduced π in logged versus primary forest in M. sepiaria (ES-π = −1.62; confidence interval-π = −2.95 to −0.3; Fig. 1). These findings were supported by analyses of inbreeding (F), which exhibited a higher signal in logged compared to primary forest in M. sepiaria (ES-π = 0.81; confidence interval-π = −0.36 to 1.99; Fig. 1, Table S2). All other species displayed comparable levels of π and F between forest types. A visualization of π ratios between logged and primary forests via Manhattan plots did not point to any substantial chromosomal regions characterized by a reduced or elevated nucleotide diversity in primary or logged forest populations (Fig. S3).

Fig. 1.

Levels of inbreeding (F coefficient; top line) and nucleotide diversity (π; bottom line) for each study species, between selectively logged and primary forests. Positive values of F indicate higher inbreeding in logged forests, and negative π values indicate lower nucleotide diversity in logged forests.

Open in a new tab

Effect sizes (ES) ± confidence intervals of inbreeding coefficient (F) and nucleotide diversity (π) between primary (unlogged) forest and selectively logged forest for each species. For F (=blue), values >0 indicate a higher signal for inbreeding in selectively logged as compared to primary forest. For π (=orange), values <0 indicate a reduced nucleotide diversity in selectively logged as compared to primary forest.

We found no significant differences in the number of runs of homozygosity (ROH) between primary and logged forest for any species (Tables S1 and S2). However, when testing for differences in ROH numbers between species regardless of forest type, we detected a significantly higher number of ROHs in M. sepiaria and Malacopteron cinereum compared to the other species, with M. sepiaria showing significantly more ROHs even than Malacop. cinereum (Fig. 2; Table S3).

Fig. 2.

Runs of homozygosity (ROH) for each of the 10 species. Each bar represents an individual, with the number of ROH reported above. Darker bars indicate birds from old-growth primary forest (UNL), while orange indicates birds from selectively logged forest (LOG).

Open in a new tab

Runs of homozygosity (ROH) for each of the ten species. Each bar represents an individual, with the number of ROH reported above. The height of each bar indicates the individual percentage of the genome covered by ROH, calculated as: [(individual sum of kb covered by ROH/genome size of the reference species) × 100]. The genome size of the reference species S. borin corresponds to 1,045,652.18 kb, while that of A. pallidus corresponds to 1,101,612.124 kb. Red indicates birds from old-growth primary forest (UNL), while orange indicates birds from selectively logged forest (LOG). Because SNPs in T. criniger were retained if genotyped in at least 70% of individuals (as compared to a 90% threshold for all other species), ROH estimates for this species may not be directly comparable and could be inflated relative to the others.

Discussion

In this study, we provide for the first time evidence of genetic erosion in a common bird species inhabiting selectively logged tropical forests. This finding adds to those reporting loss of genetic diversity in various vertebrate species following relatively recent land-use changes and other anthropogenic disturbances (Shaw et al. 2025). For example, Chattopadhyay et al. (2019) found a drastic decline in genetic diversity in a population of Sunda fruit bats (Cynopterus brachyotis) in Singapore, commencing at the onset of industrialization on the island. Overall, such results indicate that genetic erosion may occur just a few decades after the onset of habitat disturbance even in species considered common.

Loss of genetic diversity has previously been found in tropical understorey birds inhabiting forest fragments (Cros et al. 2020). In particular, insectivorous birds showed a stronger reduction in heterozygosity than frugivorous birds, probably because of their reduced dispersal capabilities. In our study, we detected a loss of nucleotide diversity and increased levels of inbreeding in M. sepiaria, a territorial insectivore with a weak flight ability. Previous research conducted in our study area showed that M. sepiaria is more likely to move shorter distances in logged forests than in pristine forest (Cosset et al. 2021). This pattern may arise because the species avoids forest gaps or perceives a higher risk of predation in selectively logged forests. Such constrained movement within logged forest interiors may limit gene flow, thereby contributing to the observed increase in inbreeding and the loss of genetic diversity. However, reduced movement in logged forests was also found in other species included in this study that do yet not show loss of genetic diversity, suggesting that movement patterns may be only one of the factors contributing to reduced genetic diversity in M. sepiaria.

Other evidence suggests that selectively logged forests constitute a low-quality habitat for M. sepiaria. The species shows significantly reduced body size in selectively logged forests compared to primary forests (Messina et al. 2021), as well as higher trophic mean level, possibly due to increased feeding on predatory arthropods (Hamer et al. 2015). Furthermore, recent findings indicate that M. sepiaria has been declining in abundance over time in both primary and selectively logged forests in our study area (Costantini et al. 2025). While in primary forests the species maintains an extensive network of co-occurrence with other understorey birds, suggesting a key role in the ecological network, in selectively logged forests its co-occurrence with other species also declines. This pattern further indicates the species’ vulnerability to both abiotic and biotic changes associated with forest selective logging.

Our analyses revealed substantial homozygosity in local populations of M. sepiaria and Malacop. cinereum, regardless of forest type. ROH are identical chromosomal segments inherited from both parents, arising more frequently with increasing parental relatedness and shared ancestry within a population (Broman and Weber 1999; Martin et al. 2023). Because our ROH analyses likely reflect historical reductions in effective population size rather than very recent inbreeding events, these results suggests that the entire local populations of M. sepiaria and Malacop. cinereum may have experienced a regional bottleneck which could have reduced their long-term adaptive potential (Barrett and Schluter 2008). This pattern is particularly pronounced in M. sepiaria, which exhibited lower genetic diversity in selectively logged forests compared to primary undisturbed forests, potentially reflecting a reduced capacity to cope with anthropogenic environmental change.

Our data show that nine out of the ten species of understorey birds, including two showing negative population abundance trends, have managed to maintain levels of genetic diversity in selectively logged forests similar to those in primary forests. These findings support the high biodiversity value of selectively logged forests for long-term conservation. However, based on species-specific population dynamics, some species may require more time to show signs of genetic erosion, keeping in mind that the most recent major logging intervention occurred only roughly two decades ago. Further studies combining long-term data on genetic variation with demographic parameters are needed to shed more light on the long-term conservation value of selectively logged forests.

As an important caveat, differences in genetic diversity among species could partly reflect phylogenetic bias. Closely related taxa, such as species within the Pellorneidae, may share similar demographic histories or life-history traits that influence levels of genetic variation. However, testing for phylogenetic signal would require a larger comparative dataset.

In conclusion, we show reduced nucleotide diversity and higher levels of inbreeding in a common rainforest songbird species, M. sepiaria, about four decades after the onset of selective logging. Furthermore, we found a substantial amount of homozygosity in the local population of M. sepiaria and Malacop. cinereum, regardless of forest type, indicating reduced standing genetic diversity. Overall, our findings raise concerns about silent population-genetic processes that might threaten the long-term persistence of these seemingly abundant species in selectively logged forests.

Materials and Methods

Study Area and Data Collection

The study area was located within the Yayasan Sabah logging concession in Sabah, Malaysian Borneo. We conducted fieldwork for avian blood collection in primary old-growth tropical forest within the Danum Valley Conservation Area (4°57045.2″N, 117°48010.4″E), and in adjacent selectively logged rainforests (Fig. 3). These lowland evergreen forests are dominated by valuable timber tree species of the family Dipterocarpaceae. Timber harvesting occurred twice: the first logging rotation (from 1976 to 1991) removed trees more than 0.6 m in diameter using tractors and high-lead cables, yielding about 120 m3 ha−1 (range: 73 to 166 m3). The second logging rotation (from 2001 to 2007) extracted trees more than 0.4 m in diameter, adding ∼31 m3 ha−1 (15 to 72 m3) of timber (Edwards et al. 2011). Following the second harvest, the heavily disturbed forest was left to recover naturally.

Fig. 3.

Geographic location of the study area, and distribution of study plots between unlogged forest (square symbols) and selectively logged forest (triangular symbols).

Open in a new tab

a) and b) Danum Valley Conservation Area in the Malaysian state of Sabah, Borneo. c) Distribution of study plots between unlogged forest (square symbols) and selectively logged forest (triangular symbols). This image was previously published in Messina et al. (2020b).

Most of the avian blood samples for DNA extraction were collected in 2023, but some in 2014. Three additional samples of T. criniger from a previous unpublished study were collected in 2017 and 2018. Fieldwork took place from early June to late August in each year. We set three plots in old-growth primary forests and three plots in twice-selectively logged forests, located at least 500 m away from the nearest road to avoid edge effect. Based on a previous study, our plots are representative of the different environmental conditions found in unlogged and selectively logged forests (Senior et al. 2018). Plots within the same type of forest were at least 1.8 km apart (mean primary forest = 6.64 km; mean logged forest = 4.04 km). Plots between primary forest and selectively logged forest were at a distance between ∼12.5 and ∼22 km (Fig. S1), ensuring that dispersal of juveniles between forest types is rare (we had no recaptures between forest types; Cosset et al. 2021). Within each plot, three parallel transects were established 250 m apart. Each transect consisted of fifteen 12-m nets set up end-to-end. All nets were operated simultaneously from 06:00 to 12:00 h on two consecutive days, after which they were relocated to the next plot. Each plot was surveyed three times per field season. Every captured bird was marked with a unique numbered ring and released after data and/or blood sample collection. Species name, ring number, day and time of capture were recorded. Blood samples (∼15 μl) were taken from the brachial vein of adult birds only using nonheparinized capillary tubes. The blood was immediately transferred into a vial and stored on ice during the time of mist netting. After transfer to the field laboratory, we added ethanol to each vial at an approximate ratio of 1:20 between blood and ethanol. The vials were then stored in a fridge.

We included in our study ten species of forest understorey birds belonging to three different families: six species of the babbler family Pellorneidae, three species of the babbler family Timaliidae, and one species of the bulbul family Pycnonotidae. Species were selected based on their capture rates and their RPA index (see below) between primary and selectively logged forest (Table 1) (Messina et al. 2020b, 2021). Babblers are strictly insectivorous and hunt for prey either on the ground or in the lower forest strata (up to 3 m), depending on species. The sole bulbul is a frugivore feeding between the understorey and midstorey levels of the forest (Wilman et al. 2014). The study species are highly sedentary and do not exhibit strong dispersal capabilities.

All experimental procedures were approved by the Sabah Biodiversity Council (access licence number: JKM/MBS.1000-2/2 JLD.16(132)). Samples were exported under the export licences JKM/MBS.1000-2/2 JLD.3(45) and JKM/MBS.1000-2/3 JLD.5(37).

RPA Index

We used species’ capture rates in the study area (same plots and transects) from 2014 to 2023, with the exception of the years 2020 and 2021 due to Covid-19 restrictions, to estimate the RPA index of each species as follows: ([abundance in logged forest—abundance in primary forest]/[abundance in logged forest + abundance in primary forest]) (Table 1). Abundance data were corrected for sampling effort (number of captures/expected effort [nets × hours]). Recaptured individuals were included in the abundance counts because recapture rates are low and similar between unlogged and logged forests (Messina et al. 2020b; Cosset et al. 2021). The RPA index is positive for species with a higher abundance in logged forest and negative for species with a higher abundance in primary forest.

Laboratory Procedures, Sequencing and Mapping

Using DNA extracted from blood samples of 96 individual birds equally subdivided between primary and logged forests, we sequenced whole genomes at an overall mean coverage of 14.2× (min = 12.77×; max = 15.81×). Additionally, three samples of T. criniger from a previous unpublished study sequenced at an average coverage of ∼10× were included in this study, adding up to a total of 99 samples (Table 1). Sample sizes were equitable across species, consisting of five to seven individuals per forest type, except for S. maculata which was represented by three individuals per forest type.

DNA was extracted using a Magnetic Animal Tissue Genomic DNA Kit (TIANGEN, Beijing, China), following the manufacturer's protocol. Genomic library preparation and genome sequencing were performed at NOVOGENE laboratory in Cambridge, United Kingdom, using paired-end sequencing on an Illumina NovaSeq X Plus PE150 platform (350-bp reads). Briefly, the genomic DNA was randomly sheared into shorter fragments, which were then end-repaired, A-tailed, and further ligated with Illumina adapters. The resulting fragments with adapters were size-selected and PCR amplified before proceeding to purification. The library was quantified through Qubit and qPCR, and its size distribution was characterized with a fragment analyzer. DNA for the three additional samples was extracted using the DNAeasy Blood and Tissue Kit (Qiagen Hilden, Germany) following the manufacturer's protocol. Libraries were prepared and then sequenced on DNBSEQ PE150 by BGI Genomics.

We assessed raw data quality of the Illumina reads using FastQC v0.12.1. We then mapped the sequences to reference genomes using the software BWA-MEM v2.2.1 and used SAMtools v1.3.1 to filter files with a minimum required MAPQ score of 20 (Li et al. 2009; Vasimuddin et al. 2019). The babbler species were mapped to a chromosome-level reference genome of Sylvia borin (Genbank assembly: GCA_014839755.1; Prost et al. 2025), the phylogenetically closest species that had a chromosome-level genome available during our study at the time (Cai et al. 2019). In contrast, the sole bulbul species was mapped to a scaffold level genome of the closely related Alophoixus pallidus, obtained from the B10K consortium (see Acknowledgements). Aligned data were then preprocessed using Picard v2.27.5 for assignment of read groups and removal of PCR duplicates.

SNP Calling and Population Genomic Analysis

To call SNPs, we used the mpileup command of BCFtools v1.17 (Li and Durbin 2009; Lefouili and Nam 2022). First, we obtained a vcf file containing all SNPs (unfiltered) and invariant sites for each species separately. Then, we used vcftools to filter out indels, low-quality SNPs (QUAL < 20), and sites for which the read depth was three times higher than the species average depth of coverage or below 3 (https://www.htslib.org/workflow/filter.html). We retained only SNPs that were genotyped in ≥90% of individuals (−max-missing 0.9), except for T. criniger for which the threshold was set at ≥70% of individuals (−max-missing 0.7) to avoid excessive loss of SNPs after filtering. This way, we obtained a filtered dataset with invariant sites and SNPs. Then, we created a vcf file containing only variant sites setting the minor allele count to 1 (−mac 1).

First, to visualize species’ genomic variation across the two populations (ie primary and logged forest) we ran PCA for each species. The PCAs were implemented on the datasets including variant sites only. Then, to estimate the number of subpopulations (K) that make up the total population in the study area for each species, we conducted admixture analyses after filtering the datasets containing variant and invariant sites (−max-missing 0.5 −maf 0.05 −thin 100,000) and using PLINK to generate bed files.

Subsequently, we calculated the inbreeding coefficient F (−het function) and ROH for each species. These analyses were implemented on the datasets including variant sites only, using PLINK v1.9 (Purcell et al. 2007). Following sensitivity analyses (Fig. S4), analyses of ROH were implemented with the flags –homozyg-density 25 –homozyg-gap 500 –homozyg-het 1 –homozyg-window-snp 50 –homozyg-window-het 5 –homozyg-window-missing 5 –homozyg-window-threshold 0.05 –homozyg-snp 50 –homozyg-kb 500 (Meyermans et al. 2020). Finally, we used the SNP dataset that included invariant sites for calculating nucleotide diversity (π) with Pixy v1.2.7 at a window size of 200 kb (Korunes and Samuk 2021).

Statistical Analysis

To assess any significant difference between primary and logged forests in inbreeding coefficient F and nucleotide diversity π, we calculated standardized Hedges’ g EF for each species. ES were calculated from means, standard deviations, and sample sizes (number of individuals) of the parameters of each species in both primary and logged forest. Following Cohen (1988), the magnitude of the ES can be considered small (Hedges g = 0.2, explaining 1% of the variance), intermediate (Hedges g = 0.5, explaining 9% of the variance) or large (Hedges g = 0.8, explaining 25% of the variance). ES are statistically significant when their confidence intervals do not overlap zero.

For each species, we calculated the ratio of π between logged and primary forest. Then, we attempted to identify loci particularly affected by a reduction in nucleotide diversity by visualizing the π ratios on Manhattan plots. The x-axis of the Manhattan plots represented the main chromosomal regions for each species except for T. criniger, whose genome had been mapped at scaffold level. Manhattan plots were implemented in the CMplot package of R (LiLin-Yin 2015).

We ran a Wilcoxon test for each species to assess differences in ROH between primary and logged forest. Finally, we ran a linear regression to assess differences between species in the number of ROH, regardless of forest type. Thus, the linear regression included the number of ROH as a dependent variable and the species as an independent variable. Pairwise comparison between species was implemented through the “emmeans” function of the homonym package. All statistical analyses were run in R.

Supplementary Material

evag139_Supplementary_Data
evag139_supplementary_data.pdf (1.5MB, pdf)

Acknowledgments

We thank the Italian Ministry of University and Research (PNRR—NextGenerationEU—Young Researchers program ID: SOE_0000189) and the European Molecular Biology Organization (EMBO Scientific Exchange Grant 10874) for supporting and funding SM. We acknowledge a Singapore National Research Foundation grant to FER (NRF-NRFI07-2021-0008) for supporting important components of this work. We thank Donald Norman, Rayzigerson Rodney Chai, and Mike Kaputa for their valuable voluntary assistance with fieldwork. We thank Fushi Ke for his support with bioinformatic analysis, Paolo Franchini for assistance with raw sequencing reads submission to NCBI, and the NUS IT's Research Computing group for computational support. We thank the B10K consortium and Guojie Zhang, Tom Gilbert, and Daniel B. Øksnebjerg for the provision of an unpublished A. pallidus genome.

Contributor Information

Simone Messina, Department of Ecological and Biological Sciences, University of Tuscia, Viterbo 01100, Italy.

David P Edwards, Department of Plant Sciences and Centre for Global Wood Security, University of Cambridge, Cambridge, UK; Conservation Research Institute, University of Cambridge, Cambridge, UK.

Suzanne Tomassi, Department of Animal and Plant Sciences, University of Sheffield, Sheffield, UK.

Suzan Benedick, Faculty of Sustainable Agriculture, Universiti Malaysia Sabah, Sandakan, Sabah 90509, Malaysia.

Yong Chee Keita Sin, Department of Biological Science, National University of Singapore, Singapore 117558, Singapore.

Frank E Rheindt, Department of Biological Science, National University of Singapore, Singapore 117558, Singapore.

David Costantini, Department of Ecological and Biological Sciences, University of Tuscia, Viterbo 01100, Italy.

Supplementary Material

Supplementary material is available at Genome Biology and Evolution online.

Conflict of Interest

The authors declare no conflict of interests.

Data Availability

Raw sequencing reads have been deposited in the NCBI Sequence Read Archive (SRA) under BioProject accession number PRJNA1450337 and are publicly available at the following link: https://www.ncbi.nlm.nih.gov/search/all/?term=PRJNA1450337. The codes used in this manuscript have been deposited in Zenodo and are publicly available at the following link: https://zenodo.org/records/20640237.

Literature Cited

  1. Barrett  R, Schluter  D. Adaptation from standing genetic variation. Trends Ecol Evol. 2008:23:38–44. 10.1016/j.tree.2007.09.008. [DOI] [PubMed] [Google Scholar]
  2. Blaser  J, Sarre  A, Poore  D, Johnson  S. Status of tropical forest management 2011. International Tropical Timber Organization; 2011. [Google Scholar]
  3. Broman  KW, Weber  JL. Long homozygous chromosomal segments in reference families from the Centre d’Étude du Polymorphisme Humain. Am J Hum Genet. 1999:65:1493–1500. 10.1086/302661. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Cai  T, et al.  Near-complete phylogeny and taxonomic revision of the world's babblers (Aves: Passeriformes). Mol Phylogenet Evol. 2019:130:346–356. 10.1016/j.ympev.2018.10.010. [DOI] [PubMed] [Google Scholar]
  5. Ceballos  G, Ehrlich  PR. Mutilation of the tree of life via mass extinction of animal genera. Proc Natl Acad Sci U S A. 2023:120:e2306987120. 10.1073/pnas.2306987120. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Chattopadhyay  B, Garg  KM, Mendenhall  IH, Rheindt  FE. Historic DNA reveals anthropocene threat to a tropical urban fruit bat. Curr Biol. 2019:29:R1299–R1300. 10.1016/j.cub.2019.11.013. [DOI] [PubMed] [Google Scholar]
  7. Cohen  J. Statistical power analysis for the behavioral sciences. 2nd ed. L. Erlbaum Associates; 1988. [Google Scholar]
  8. Cosset  CCP, et al.  Impacts of tropical selective logging on local-scale movements of understory birds. Biol Conserv. 2021:264:109374. 10.1016/j.biocon.2021.109374. [DOI] [Google Scholar]
  9. Costantini  D, et al.  Selective logging weakly influences species co-occurrence in a community of tropical understorey birds. J Anim Ecol. 2025:94:1757–1769. 10.1111/1365-2656.70085. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Costantini  D, Edwards  DP, Simons  MJP. Life after logging in tropical forests of Borneo: a meta-analysis. Biol Conserv. 2016:196:182–188. 10.1016/j.biocon.2016.02.020. [DOI] [Google Scholar]
  11. Cros  E, et al.  Fine-scale barriers to connectivity across a fragmented South-East Asian landscape in six songbird species. Evol Appl. 2020:13:1026–1036. 10.1111/eva.12918. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. DeWoody  JA, Harder  AM, Mathur  S, Willoughby  JR. The long-standing significance of genetic diversity in conservation. Mol Ecol. 2021:30:4147–4154. 10.1111/mec.16051. [DOI] [PubMed] [Google Scholar]
  13. Dussex  N, Morales  HE, Grossen  C, Dalén  L, Van Oosterhout  C. Purging and accumulation of genetic load in conservation. Trends Ecol Evol. 2023:38:961–969. 10.1016/j.tree.2023.05.008. [DOI] [PubMed] [Google Scholar]
  14. Eaton  JA, et al.  Birds of the Indonesian archipelago: greater sundas and wallacea. 2nd ed. Lynx; 2021. [Google Scholar]
  15. Edwards  DP, et al.  Degraded lands worth protecting: the biological importance of Southeast Asia's repeatedly logged forests. Proc Biol Sci. 2011:278:82–90. 10.1098/rspb.2010.1062. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. FAO . Global forest resources assessment 2025. FAO; 2025. [Google Scholar]
  17. Figueiredo  L, Krauss  J, Steffan-Dewenter  I, Sarmento Cabral  J. Understanding extinction debts: spatio-temporal scales, mechanisms and a roadmap for future research. Ecography. 2019:42:1973–1990. 10.1111/ecog.04740. [DOI] [Google Scholar]
  18. Gargiulo  R, Budde  KB, Heuertz  M. Mind the lag: understanding genetic extinction debt for conservation. Trends Ecol Evol. 2024:40:228–237. 10.1016/j.tree.2024.10.008. [DOI] [PubMed] [Google Scholar]
  19. Hamer  KC, et al.  Impacts of selective logging on insectivorous birds in Borneo: the importance of trophic position, body size and foraging height. Biol Conserv. 2015:188:82–88. 10.1016/j.biocon.2014.09.026. [DOI] [Google Scholar]
  20. Imai  N, Sugau  JB, Pereira  JT, Titin  J, Kitayama  K. Impacts of selective logging on spatial structure of tree species composition in Bornean tropical rain forests. J For Res. 2019:24:335–340. 10.1080/13416979.2019.1666958. [DOI] [Google Scholar]
  21. Korunes  KL, Samuk  K. Pixy: unbiased estimation of nucleotide diversity and divergence in the presence of missing data. Mol Ecol Resour. 2021:21:1359–1368. 10.1111/1755-0998.13326. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Kuussaari  M, et al.  Extinction debt: a challenge for biodiversity conservation. Trends Ecol Evol. 2009:24:564–571. 10.1016/j.tree.2009.04.011. [DOI] [PubMed] [Google Scholar]
  23. Lefouili  M, Nam  K. The evaluation of Bcftools mpileup and GATK HaplotypeCaller for variant calling in non-human species. Sci Rep. 2022:12:11331. 10.1038/s41598-022-15563-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Li  H, et al.  1000 genome project data processing subgroup. The sequence alignment/map format and SAMtools. Bioinformatics. 2009:25:2078–2079. 10.1093/bioinformatics/btp352. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Li  H, Durbin  R. Fast and accurate short read alignment with Burrows–Wheeler transform. Bioinformatics. 2009:25:1754–1760. 10.1093/bioinformatics/btp324. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. LiLin-Yin . 2015. CMplot: Circle Manhattan Plot. R package version 4.5.1. https://CRAN.R-project.org/package=CMplot
  27. Martin  CA, et al.  Runs of homozygosity reveal past bottlenecks and contemporary inbreeding across diverging populations of an island-colonizing bird. Mol Ecol. 2023:32:1972–1989. 10.1111/mec.16865. [DOI] [PubMed] [Google Scholar]
  28. Messina  S, et al.  Impacts of selective logging on the oxidative status of tropical understorey birds. J Anim Ecol. 2020a:89:2222–2234. 10.1111/1365-2656.13280. [DOI] [PubMed] [Google Scholar]
  29. Messina  S, et al.  Glucocorticoids link forest type to local abundance in tropical birds. Funct Ecol. 2020b:34:1814–1825. 10.1111/1365-2435.13586. [DOI] [Google Scholar]
  30. Messina  S, et al.  Selective logging reduces body size in omnivorous and frugivorous tropical forest birds. Biol Conserv. 2021:256:109036. 10.1016/j.biocon.2021.109036. [DOI] [Google Scholar]
  31. Meyermans  R, Gorssen  W, Buys  N, Janssens  S. How to study runs of homozygosity using PLINK? A guide for analyzing medium density SNP data in livestock and pet species. BMC Genomics. 2020:21:94. 10.1186/s12864-020-6463-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Paez  S, et al.  Reference genomes for conservation. Science. 2022:377:364–366. 10.1126/science.abm8127. [DOI] [PubMed] [Google Scholar]
  33. Prost  S, et al.  The unexpected loss of the ‘hunger hormone’ ghrelin in true passerines: a game changer in migration physiology. R Soc Open Sci. 2025:12:242107. 10.1098/rsos.242107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Purcell  S, et al.  PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet. 2007:81:559–575. 10.1086/519795. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Putz  FE, et al.  Sustaining conservation values in selectively logged tropical forests: the attained and the attainable. Conserv Lett. 2012:5:296–303. 10.1111/j.1755-263X.2012.00242.x. [DOI] [Google Scholar]
  36. Schmidt  C, Hoban  S, Hunter  M, Paz-Vinas  I, Garroway  CJ. Genetic diversity and IUCN Red List status. Conserv Biol. 2023:37:e14064. 10.1111/cobi.14064. [DOI] [PubMed] [Google Scholar]
  37. Senior  RA, Hill  JK, Benedick  S, Edwards  DP. Tropical forests are thermally buffered despite intensive selective logging. Glob Chang Biol. 2018:24:1267–1278. 10.1111/gcb.13914. [DOI] [PubMed] [Google Scholar]
  38. Shaw  RE, et al.  Global meta-analysis shows action is needed to halt genetic diversity loss. Nature. 2025:638:704–710. 10.1038/s41586-024-08458-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Vasimuddin  MD, Misra  S, Li  H, Aluru  S. 2019. Efficient architecture-aware acceleration of BWA-MEM for multicore systems. In: 2019 IEEE International Parallel and Distributed Processing Symposium (IPDPS). Rio de Janeiro, Brazil: IEEE. p. 314–324.
  40. Wilman  H, et al.  EltonTraits 1.0: species-level foraging attributes of the world's birds and mammals: ecological archives E095-178. Ecology. 2014:95:2027. 10.1890/13-1917.1. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

evag139_Supplementary_Data
evag139_supplementary_data.pdf (1.5MB, pdf)

Data Availability Statement

Raw sequencing reads have been deposited in the NCBI Sequence Read Archive (SRA) under BioProject accession number PRJNA1450337 and are publicly available at the following link: https://www.ncbi.nlm.nih.gov/search/all/?term=PRJNA1450337. The codes used in this manuscript have been deposited in Zenodo and are publicly available at the following link: https://zenodo.org/records/20640237.

Articles from Genome Biology and Evolution are provided here courtesy of Oxford University Press

RESOURCES