Duplicate of 10.5061/dryad.6hdr7sqxq

Human activities are primarily responsible for habitat loss and changes in natural environments around the world. It has been suggested that populations inhabiting human-modified landscapes experience reduced gene flow, inbreeding depression, and loss of alleles due to genetic drift. However, the empirical evidence shows contrasting effects of anthropogenic disturbances on the genetic diversity of species. We performed a meta-analysis of 61 studies that compared the genetic diversity of plant and/or animal populations in disturbed and more preserved areas (316 paired comparisons) to investigate the genetic responses to different disturbance types. There is a negative effect (effect size: -0.45; 95% CI: -0.61, -0.29) of disturbances on genetic diversity, in which the most detrimental effects are caused by the loss of connectivity and forest cover. The methodological approach can explain part of the heterogeneity among the genetic responses detected by primary studies: (i) studies using the number of effective alleles did not detect genetic erosion, while all other indices, revealed negative responses to disturbances; and (ii) only studies performed with transferred or a combination of transferred and specific microsatellites detected negative responses. The effects on animal populations are more detrimental than in plant populations. Only plant species with shrub life form, self-incompatible reproductive systems, and biotic pollination and seed dispersal, showed negative responses to disturbances. Despite all heterogeneity among studies, there is an overall negative effect of disturbances on the genetic diversity, which indicates that remaining populations inhabiting human-modified landscapes have reduced evolutionary potential and are prone to local extinction.

We searched for studies that evaluated the effects of anthropogenic disturbances on plant and animal genetic diversity published before February 07^th, 2019. The search was performed in the Scopus database using the following term sequences: [("Genetic diversity") AND (Microsatellite OR SSR) AND ("fragmented habitat" OR "habitat fragmentation" OR deforest* OR logging OR "habitat disturbance" OR "disturbed habitat" OR "habitat loss")]. These terms were searched for in the title, abstract, and keyword sections of the manuscripts, except for the sequence of terms related to the genetic marker, which were searched for in the entire manuscript.

Initially, we identified 1,263 manuscripts that were screened based on the following additive criteria (Figure S1, Supporting information 1): I. studies that explicitly evaluated the effects of anthropogenic disturbances on the genetic diversity of terrestrial plants and/or animals. We considered six genetic descriptive parameters as measures of genetic diversity: observed and expected heterozygosity (H_O and H_E), allelic richness (Ar), mean number of alleles (A), mean number of effective alleles (Ae), and mean number of private alleles (Ap); II. studies that used microsatellite markers; III. studies that evaluated genetic diversity along a gradient of anthropogenic disturbances or compared it between relatively preserved (control) and disturbed (treatment) areas in the same geographical region. We considered four categories of local-scale disturbance – (1) habitat quality, (2) patch size, (3) edge effects, and (4) land use intensification – and three categories of landscape-scale disturbance – (1) landscape quality, (2) connectivity, and (3) forest cover; IV. studies that provided values of sample size and genetic parameters (mean estimate and error measure for each treatment, or linear regression coefficients for gradient studies) for all evaluated populations; V. studies with replication of treatments and control groups, and VI. studies published in the English language. Details about the criteria for studies inclusion are presented in the Supplementary Information. The reviews and meta-analyses identified in the initial search were used as a source of additional papers for our database, to which 4 additional manuscripts were included.

This screening procedure resulted in 61 studies and 316 paired comparisons (mean ± SD = 5.2 ± 3.1 comparisons per study). For each paired comparison we extracted the sample size, mean value of genetic parameter and standard deviation per treatment. In case of gradient studies, we extracted the sample size and the linear regression coefficient. Additionally, we extracted information about each study methodology such as the genetic parameter, type of microsatellite marker used, number of loci analyzed, investigated taxa (amphibians, birds, insects, mammals, reptiles, and plants), and life history traits of plant species . Whenever information was not available in the manuscript, we searched the literature to obtain it.