Non-Invasive Sampling for Population Genetics of Wild Terrestrial Mammals (2015–2025): A Systematic Review

Abstract

Genetic variability in terrestrial mammals is essential for understanding population and evolutionary dynamics, as well as for establishing effective strategies in conservation biology. This comprehensive review aimed to critically analyze invasive and non-invasive techniques used to assess genetic variability in wild terrestrial mammals. Using the PICO (Population, Intervention, Comparison, Outcome) format and following PRISMA guidelines, a comprehensive literature search was conducted in Web of Science, Scopus and Science Direct databases, including articles published in English from January 2015 to April 2025. Thirty-one experimental studies were selected that met specific criteria related to genetic evaluation using invasive (direct blood or tissue collection) and non-invasive (stool, hair and saliva collection) techniques. The results indicate that invasive techniques provide samples of high genetic quality, albeit with important ethical and animal welfare considerations. In contrast, non-invasive techniques offer less disruptive methods, although they present significant challenges in terms of quantity and purity of DNA obtained, potentially affecting the accuracy and confidence of genetic analysis. Detailed analysis of selected studies showed diverse patterns of heterozygosity and inbreeding coefficients between different taxonomic orders (Carnivora, Artiodactyla, Proboscidea, Primates and Rodentia). In addition, the main anthropogenic threats and current conservation strategies implemented in different species were identified. An overall genetic variability ranging from high to moderate was observed, with large species being more vulnerable to genetic reduction due to changes in habitat and human activities. Rather than a static comparison, our synthesis traces a clear methodological arc from small short tandem repeats (STR, or microsatellites) panels towards SNP-based approaches enabled by next-generation sequencing, including reduced representation (ddRAD), amplicon panels (GT-seq), and hybridisation capture tailored to degraded DNA from hair, faeces, and environmental substrates. Over 2015–2025, study designs shifted from presence/absence and coarse diversity estimates to robust inference of relatedness, assignment, effective population size, and gene flow using hundreds–thousands of SNPs and genotype-likelihood frameworks tolerant of allelic dropout and low coverage. Laboratory practice converged on multi-tube replication, synthetic blocking oligos, and capture-based enrichment; bioinformatics adopted probabilistic genotype calling, error-aware filtering, and replication-based consensus. This review provides a solid basis for optimizing genetic sampling methods, allowing for more ethical and efficient studies. Furthermore, it contributes to strengthening conservation strategies by underlining the importance of adapting the sampling method to the biological and ecological particularities of each species studied. Ultimately, these findings can significantly improve genetic conservation decision-making, benefiting the sustainability and resilience of wild land mammal populations.

1. Introduction

Genetic variability in terrestrial mammals is a central element for understanding the ecological and evolutionary dynamics of populations, as well as for the effective implementation of biological conservation measures. This variability refers to the observable genetic diversity within populations, resulting mainly from mutational processes and genetic recombination, which generate new alleles or redistribute existing ones [1]. Such processes can be significantly influenced by environmental and anthropogenic factors, which modify both the structure and long-term viability of populations [2].

Accurate estimation of genetic diversity relies in the use of advanced tools such as molecular markers including short tandem repeats (STR, or microsatellites), single nucleotide polymorphisms (SNPs) and copy number variants (CNVs). These tools provide detailed information on structural and point variations in DNA, allowing critical parameters such as heterozygosity and inbreeding coefficients to be assessed [3]. High genetic diversity provides greater resilience to disease, environmental stresses and anthropogenic changes, while low diversity can compromise the survival and viability of wild populations [4]. Genetic analysis in terrestrial mammals requires the collection of biological samples, which can be collected using invasive and non-invasive techniques. Invasive techniques such as direct blood or tissue collection, while providing high quality and quantity of DNA, raise important ethical and animal welfare considerations [5].

In contrast, non-invasive techniques, such as faecal, hair and saliva collection, minimize stress on the individuals studied, although they face challenges in terms of DNA quality and quantity, which may affect the accuracy of genetic analysis [6]. In this context, the present systematic review aims to critically analyze the invasive and non-invasive techniques currently used to assess genetic variability in terrestrial mammals. Through this review we aim to clearly identify the advantages and limitations of each technique, highlighting their applicability in genetic studies and in decision-making for conservation and population management. This review will contribute significantly to the advancement of conservation biology by providing a solid basis for optimizing genetic sampling methodologies, promoting more ethical and effective studies, and strengthening conservation strategies based on genetic science.

2. Methodology

2.1. Literature Search

In the present comprehensive review, we used the Population (P), Intervention (I), Comparison (C) and Outcome (O) format proposed by [7] to formulate the research question. Briefly, P was mammalian animals, I was genetic variability, C was the use of invasive and non-invasive techniques for tissue procurement, and O was the reported information on heterozygosity, inbreeding coefficients, and level of genetic variability in terrestrial mammalian animals. The present systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [8]. Eligible published scientific articles were identified through Web of Science, Scopus and ScienceDirect databases. Boolean operators (AND and OR) were used to search for articles. The keyword string syntax was (“non-invasive sampling” OR “invasive sampling” OR “animal tissue”) AND (“mammals” OR “wildlife” OR “wild animals”) AND (“genetic variability” OR “inbreeding coefficients” OR “heterozygosity”). For up-to-date information, this systematic review only considered studies published between January 2014 and April 2025 and written in English.

2.2. Inclusion and Exclusion Criteria

Only scientific articles that met the following inclusion criteria were considered in this systematic review: (1) experimental studies assessing genetic variability in terrestrial wild mammals with invasive and non-invasive sampling; (2) articles published in English in peer-reviewed scientific journals between January 2014 and April 2025; and (3) studies providing data on genetic variability in terrestrial wild mammals with invasive and non-invasive sampling. During the study selection process, manuscripts published before 2002, not written in English, or using animals other than terrestrial free-living mammals were excluded. Duplicate papers, scientific conference abstracts, theses, books or book chapters, and traditional narrative review articles were also excluded. A total of 31 scientific articles met the inclusion criteria and were analyzed in this review (Figure 1). Relevant information was extracted from each study, including the title, authors, abstract, year of publication and location of the study, and compiled into an Excel spreadsheet. The spreadsheet also included all the key data needed to address the research question comprehensively.

3. Invasive and Non-Invasive Genetic Sampling in Terrestrial Mammals: Techniques, Challenges, and Conservation Applications

Invasive and non-invasive genetic sampling constitute the two principal methodologies employed in wildlife conservation genetics (Figure 2). Each approach uses distinct techniques and therefore exhibits specific advantages and limitations that must be acknowledged when estimating the genetic variability of wild terrestrial mammals [9].

The technique of darting has been the subject of debate in scientific discourse, with its classification as either invasive or non-invasive being contingent upon the definition of invasiveness. In the present study, darting is regarded as invasive and will not be incorporated within this category. This is in accordance with the widely accepted definition of “non-invasive” in the field of population genetics (i.e., no capture, manipulation, or skin puncture of the animal in question); remote biopsy with a dart is not non-invasive.

Historically, invasive sampling—typically involving capture, anaesthesia, and tissue biopsies—dominated conservation genetics. Although it often yields high-quality DNA, the technique is costly, labour-intensive, and can impose severe stress or injury on animals, thereby contradicting modern ethical standards that prioritise less intrusive methods [9,10]. Over the past two decades, non-invasive genetic sampling has gained prominence as an ethically preferable and logistically flexible alternative (

Table 1. Summary of published studies in which both non-invasive and invasive genetic sampling were conducted in terrestrial mammals.

Species	N	n of Non-Invasive Samples	n of Invasive Samples	Type of Marker	Reference
Otocolobus manul	16	6	10	STR,12S ribosomal	[12]
Lynx lynx	6	5	1	STR	[13]
Puma concolor	3355	0	3355	STR	[14]
Ursus arctos	44	27	17	STR	[15]
Lutra lutra	47	47	0	STR, mitochondrial cytochrome b gene	[16]
Panthera uncia	213	213	0	STR, mitochondrial cytochrome b gene	[17]
Leopardus geoffroyi	172	163	9	STR	[18]
Mellivora capensis	194	113	81	STR	[19]
Panthera onca	536	536	0	STR	[20]
Panthera tigris	718	718	0	STR	[21]
Vulpes vulpes	150	150	0	STR	[22]
Felis nigripes	23	23	0	STR, mithochondrial DNA ND5	[23]
Panthera tigris tigris	71	12	59	STR	[24]
Alces alces	1058	529	529	STR	[25]
Ozotoceros bezoarticus	155	142	13	STR, mithochondrial DNA	[26]
Odocoileus hemionus fuliginatus	238	238	0	STR	[27]
Cervus hanglu hanglu	160	160	0	STR, mitochondrial D-loop	[28]
Elephas maximus	252	252	0	STR, mitochondrial DNA	[29]
Alouatta caraya	138	132	6	STR, mitochondrial DNA	[30]
Sciurus griseus	117	114	3	STR, mitochondrial DNA D-loop	[31]
Hydrochoerus hydrochaeris	54	27	27	STR, mitochondrial DNA	[32]
Loxodonta africana	37	37	0	Mitochondrial genomes	[33]
Tremarctos ornatus	38	38	0	STR, mitochondrial D-loop	[34]

). Faecal sampling is now the most widely applied non-invasive source because it is easy to collect, logistically inexpensive, and does not require animal handling; nevertheless, it demands multiple replicates to secure reliable genotypes owing to low DNA quantity [11].

As expected, invasive sampling achieved the highest success rate across all technologies (Table 1). There was a significant improvement associated with SNP-based methods relative to STRs in non-invasive matrices, as well as a smaller but consistent advantage of invasive over non-invasive sampling. Capture- or amplicon-based SNP panels have higher initial costs but lower marginal costs when scaled up to 96–192 samples, resulting in competitive or superior costs per sample that pass quality control (QC) compared with STRs once batch sizes increase.

Hair represents another common non-invasive substrate that generally contains higher endogenous DNA concentrations than faeces; however, retrieving and amplifying hair DNA still calls for careful extraction protocols to minimise allelic dropout and genotyping errors [35]. Saliva and chewed plant material offer unique, minimally disruptive opportunities for collecting buccal epithelial cells that supply high-quality nuclear DNA; nevertheless, stringent field and laboratory protocols are required to prevent contamination and DNA degradation when using these matrices in population-genetic analyses [33].

Environmental DNA (eDNA) has emerged as a cutting-edge methodology in conservation genetics. By analysing the DNA shed into soils, sediments or bodies of water, researchers can infer the presence of species and genetic diversity without direct observation, provided rigorous contamination control and taxonomic verification standards are met [36,37]. The preservation of collected samples is critical, particularly for non-invasive materials which often contain fragmented or inhibitor-laden DNA. Environmental DNA (eDNA) metabarcoding is a non-invasive method of surveying biodiversity that analyses extracellular and shed DNA present in environmental samples (e.g., water, soil, air). After sampling, the DNA is extracted, and short taxonomically informative loci (commonly the mitochondrial 12S or COI gene for animals and the 16S gene for prokaryotes) are amplified by PCR and sequenced using high-throughput platforms. Bioinformatic pipelines then assign sequences to taxa using curated reference databases, producing multi-species inventories from a single sample. Silica gel desiccation, cold-chain transport, and the addition of chelating buffers remain the most reliable strategies for retaining genetic integrity and thereby reducing downstream analytical error. When properly preserved, non-invasive substrates permit the successful amplification of both mitochondrial (mtDNA) and nuclear (nDNA) markers, each providing complementary information on matrilineal history and contemporary gene flow, respectively [9]. Mitochondrial DNA is frequently chosen because of its high copy number per cell, rapid evolutionary rate, and maternal mode of inheritance, which collectively facilitate inferences on phylogeography and historical demography in mammalian populations [38,39].

Among nuclear loci, STR remain the most widely applied markers in conservation-genetic studies owing to their high polymorphism and codominant inheritance. STR panels developed for carnivores and ungulates have revealed substantial genetic diversity and have served to identify incipient inbreeding in isolated populations. Sex-linked markers—in particular those located on the Y-chromosome and the X-inactivation centre—are essential for diagnosing sex ratio, paternity, and dispersal patterns, facilitating the design of sex-specific management interventions [40]. Despite these advantages, molecular markers derived from non-invasive samples are susceptible to genotyping errors caused by allelic dropout, false alleles, and heterozygote deficiency, all of which can bias population-genetic parameters if left uncorrected [41]. DNA degradation constitutes a major challenge because environmental exposure accelerates depurination and oxidative damage, producing short fragments that are difficult to amplify. Carnivore faeces often contain higher concentrations of digestive inhibitors than herbivore dung owing to their protein-rich diet and gut physiology [42].

Low DNA yield represents a further constraint, particularly when studying elusive or endangered species with inherently small population sizes. Studies on felids and mustelids have routinely reported amplification failure rates exceeding 50% due to marginal DNA quality and quantity [43]. Plant secondary metabolites, humic acids, and bile salts act as potent polymerase inhibitors in faecal DNA extracts, significantly diminishing amplification efficiency [44]. Several technical solutions have been proposed to improve DNA recovery from challenging substrates, including the adoption of combination extraction kits, inhibitor-binding resins, and low-temperature lysis protocols, all of which have markedly enhanced genotyping success [45]. External PCR controls, allele-specific replication, and the adoption of multi-tube consensus approaches effectively mitigate amplification artefacts associated with fragmented DNA templates [46]. High-throughput sequencing technology, coupled with rigorous replication and stochastic error-modelling frameworks, has substantially reduced genotyping error rates and improved individual identification accuracy [9,35]. In summary, although non-invasive genetic sampling introduces technical obstacles, recent methodological advances have minimized these limitations, making the approach a robust, ethical, and cost-effective tool for monitoring genetic diversity, population structure, and demographic trends in wild terrestrial mammals, thereby providing indispensable data for evidence-based conservation.

4. Genetic Variability in the Order Carnivora

4.1. Main Species and Their Heterozygosity Values

The Carnivoraformes clade comprises mammals such as tigers, lions, bears, raccoons and others including their ancestors, distinguished by their diet and jaw structure as mesocarnivores (50% to 70% animal matter), hypocarnivores (<50% animal material), hypercarnivorous (>70% animal matter) or bone breakers [47]. At least for this group, studies of autosomal heterozygosity are confirmed in felids such as Iberian bobcats, pumas, cheetahs, jaguars, lions and leopards [48]. The trend in their genetic diversity is related to the dimensionality of their populations for microsatellite studies [49]. In addition, it is known that the decline in their populations causes high inbreeding values, reducing the gene pool of the species, mainly due to anthropogenic causes [50]. In this regard,

Table 2. Main species studied and their genetic heterogeneity values in the order Carnivora.

Species	Main Threats	Conservation Strategy	Type of Sample	Expected and Observed Heterocigosity *	Reference
Otocolobus manul	Habitat degradation and climate change	Poaching mitigation, creation of NPAs and Captive Areas	Faeces and blood.	He = 0.62 Ho = 0.57	[12]
Lynx lynx	Poaching and poor habitat quality	Reintroduction	Faeces, blood and tissue	-	[13]
Puma concolor	Urbanisation and habitat loss and fragmentation	Protected natural areas	Skin and muscle	He = 0.59 Ho = 0.52	[14]
Ursus arctos arctos	Human activities	Reintroduction	Blood, tissue, excreta and hair	He = 0.61 Ho = 0.64	[15]
Ursus arctos marsicanus	Human activities	Reintroduction	Blood, tissue, excreta and hair	He = 0.40 Ho = 0.39	[15]
Lutra lutra	Habitat fragmentation	Reintroduction	Faeces	Ho = 0.37	[16]
Panthera uncia	Habitat fragmentation	Translocation	Faeces	He = 0.57 Ho = 0.54	[17]
Leopardus geoffroyi	Habitat fragmentation and land use change	Biological corridors	Tissue and blood	He = 0.77 Ho = 0.70	[18]
Meles meles	Urbanisation	Biological corridors	Fresh, dead tissue and faeces	He = 0.45 Ho = 0.42	[19]
Panthera onca	Habitat fragmentation	Biological corridors and protected areas	Faeces	He = 0.61 Ho = 0.55	[20]
Panthera tigris tigris	Poaching and habitat loss	Biological corridors	Faeces	He = 0.77 Ho = 0.68	[21]
Vulpes vulpes	Urbanisation and population isolation	-	Faeces	He = 0.55 Ho = 0.52	[22]
Felis nigripes	Overgrazing and extensive agriculture	Captive areas	Hair	He = 0.62 Ho = 0.68	[23]
Panthera tigris tigris	Poaching and habitat loss	Biological corridors	Blood and excreta	He = 0.64 Ho = 0.50	[24]

* In some cases, this is average heterozygosity across populations, when more than two populations are involved.

shows studies related to expected and observed heterozygosity in the order Carnivora and their inbreeding pairings, where it is indicated that when the value of Ho is less than He, it is possible that more inbred matings exist. In addition, each of the threats to the populations and their implemented conservation strategies are described. Furthermore, their genetic variability status according to the authors is included, using expected heterozygosity (He) thresholds to categorize diversity as low, moderate, or high.

4.2. Population Balance and Inbreeding Coefficients

The Hardy–Weinberg equilibrium is commonly used for population analysis in mammals as an input to genetic analyses using STR that can be performed in free-living and captive species. STR can also be used to identify null alleles and heterozygosity deficits in subpopulations, providing the framework for evaluating deviations from equilibrium and quantifying Wright’s F-statistics [14,24]. In populations, loci in Hardy–Weinberg equilibrium have heterozygosity levels that conform to expected values, but inbreeding coefficient Fis varies from −1 to +1 values. For example, in the study conducted on the species Otocolobus manul, genetic diversity values ranged from low to moderate levels with Fis values ranging from −0.093 to 0.105 for zoo-managed and wild populations, respectively [12]. In another similar case in Puma concolor, almost all loci were under the null expectation of HWE, with an observed heterozygosity significantly lower than expected heterozygosity and Fis ranging from 0.05 to 0.12 with a mean of 0.07; Felis nigripes was estimated at 0.048, both with moderate diversity [14,23]. However, an imbalance in the equilibrium can reveal biological and ecological patterns, as in the case of Vulpes vulpes where there was HWE imbalance indicating a reduced number of individuals, absence of random mating, and inbreeding, with an approximate value of Fis = 0.047 [22]. A similar pattern was observed in the case of Lynx lynx, where genetic differentiation among populations had lower values Fst = 0.097 [13].

Positive Fis values also herald a deviation in HWE and heterozygote deficiency, as was the case for Meles meles, with a value at Fis = 0.021 [19]. Furthermore, deviation from equilibrium can be significant in some species that maintain stable genetic variability in their populations; this apparent contradiction arises when the expected heterozygosity (He) remains moderate or high while observed heterozygosity (Ho) does not, such as in the cases of Leopardus geoffroyi, with a population value of Fis = 0.093, and Panthera leo, Fis = 0.03 [18,51]. It is possible to have this value in a single locus, as suggested by [21] in the species Panthera tigris tigris with an estimated value Fis = 0.208.

In summary, based on available data, certain species within the carnivorous order can maintain high-moderate variability with different mammal conservation strategies, mostly reintroduction programmes, biological corridors and establishment of natural protected areas. Anthropogenic adversities have threatened some species; however, those that maintain moderate variability are latently deprived if more conservation measures appropriate to each situation are not implemented. Most of the genetic studies were carried out with STR considering the Hardy–Weinberg equilibrium law, estimating observed and expected heterozygosity as well as inbreeding coefficients, and were based mostly on DNA extracted from invasive and noninvasive samples. Such tools are fundamental for decision-making in carnivore conservation in different parts of the world to increase species variability and coexistence of populations.

5. Genetic Variability in the Order Artiodactyla

5.1. Main Species and Their Heterozygosity Values

The main characteristic of artiodactyls is that they have a pair of toes or digits, like ungulate mammals and other paraxonians [52]. In this order, genetic variability studies have been carried out measuring heterozygosity [53]. Some authors consider that the size of larger and more mobile mammals has lower levels of genetic variability than that of smaller and less mobile mammals, which implies the search for new habitat territories [54], hence the importance of estimating their variation. Table 3 shows the expected and observed heterozygosity values, threats and conservation strategies for 10 species belonging to the order Artiodactyla.

5.2. Population Balance and Inbreeding Coefficients

Different studies on ungulates estimated the Hardy–Weinberg equilibrium and inbreeding coefficients. For instance, the species Ozotoceros bezoarticus had an estimated Fis coefficient of 0.157, which could suggest a slight difference in heterozygotes and indicate a certain level of inbreeding [26]. Similarly, in the species Cervus elaphus hanglu, which experienced a genetic bottleneck that altered mutations and provided high heterozygosity values with respect to HWE equilibrium, the Fis coefficient was estimated at 0.38 [28]. By contrast, population deviation in Sus scrofa indicated the Wahlund effect, whereby the population substructure prevented free mating and reduced heterozygosity, with Fis estimated at 0.051 [55].

Table 3. Main species studied and their heterogeneity values in the order Artiodactyla.

Species	Main Threats	Conservation Strategy	Genetic Sampling	Expected and Observed Heterocigosity	Reference
Ammotragus lervia	Illegal hunting	Conservation and translocation management units	Faeces, hair, bone and tissue	He = 0.49 Ho = 0.43	[56]
Ammotragus lervia	Poaching and habitat loss	None	Blood and tissue	He = 0.48 Ho = 0.46	[57]
Alces americanus americanus	Physical and anthropogenic barriers	Connectivity between towns	Hair and fabric	He = 0.32 Ho = 0.35	[25]
Sus scrofa	Habitat reduction	Urban colonisation	Muscle and tissue	He = 0.59 Ho = 0.56	[55]
Naemorhedus griseus	Consanguinity	Species in captivity for reintroduction	Blood	He = 0.45 Ho = 0.19	[58]
Lama glama	Species utilisation	Species in captivity for reintroduction	Blood	He = 0.76 Ho = 0.71	[59]
Camelus dromedarius	None	Research only	Blood	He = 0.72 Ho = 0.67	[60]
Ozotoceros bezoarticus	Agriculture and urbanisation	Population genetic variability studies	Faeces	He = 0.75 Ho = 0.72	[26]
Tragelaphus eurycerus ssp. isaaci	Human activities	Species in captivity for reintroduction	Faeces and tissue	He = 0.42 Ho = 0.71	[61]
Odocoileus hemionus fuliginatus	Fragmentation of habitat and urbanisation	Natural protected areas and biological corridors	Faeces	He = 0.56 Ho = 0.56	[27]
Cervus elaphus hanglu	Habitat reduction	Protected natural areas	Hair	He = 0.66 Ho = 0.44	[28]

Table 3. Main species studied and their heterogeneity values in the order Artiodactyla.

Species	Main Threats	Conservation Strategy	Genetic Sampling	Expected and Observed Heterocigosity	Reference
Ammotragus lervia	Illegal hunting	Conservation and translocation management units	Faeces, hair, bone and tissue	He = 0.49 Ho = 0.43	[56]
Ammotragus lervia	Poaching and habitat loss	None	Blood and tissue	He = 0.48 Ho = 0.46	[57]
Alces americanus americanus	Physical and anthropogenic barriers	Connectivity between towns	Hair and fabric	He = 0.32 Ho = 0.35	[25]
Sus scrofa	Habitat reduction	Urban colonisation	Muscle and tissue	He = 0.59 Ho = 0.56	[55]
Naemorhedus griseus	Consanguinity	Species in captivity for reintroduction	Blood	He = 0.45 Ho = 0.19	[58]
Lama glama	Species utilisation	Species in captivity for reintroduction	Blood	He = 0.76 Ho = 0.71	[59]
Camelus dromedarius	None	Research only	Blood	He = 0.72 Ho = 0.67	[60]
Ozotoceros bezoarticus	Agriculture and urbanisation	Population genetic variability studies	Faeces	He = 0.75 Ho = 0.72	[26]
Tragelaphus eurycerus ssp. isaaci	Human activities	Species in captivity for reintroduction	Faeces and tissue	He = 0.42 Ho = 0.71	[61]
Odocoileus hemionus fuliginatus	Fragmentation of habitat and urbanisation	Natural protected areas and biological corridors	Faeces	He = 0.56 Ho = 0.56	[27]
Cervus elaphus hanglu	Habitat reduction	Protected natural areas	Hair	He = 0.66 Ho = 0.44	[28]

In particular, the order Artiodactyla is considered important because of the different uses of the species that comprise it, most of which are used for poaching, which if controlled legally does not affect the populations; however, when there is no control, the populations tend to decline, and this affects the genetic variability of the species. It is worth mentioning that some species exhibit invasive behaviour that may affect other species in their niche; despite this, most are in some category of threat. Compared to other orders, studies were mostly carried out with invasive samples and a minority with faecal samples; this encouraged the advancement of genetic sampling and heterozygosity–endogamy analysis for populations. In general, the order maintains moderate-low genetic variability, due to anthropogenic causes.

6. Genetic Variability in the Orders Proboscidea, Primates and Rodentia

6.1. Main Species and Their Expected and Observed Heterozygosity Values

The order Proboscidea comprises large mammals such as living and extinct elephants (Elephantidae), characterised by the proboscis (trunk), which is a combination of nose and upper lip, and enlarged tusks. Genetic variability studies have differentiated elephant species, adding that species in a forested habitat are more variable than in the African savannah [62]. In another case, the order Primates is divided into four genera Homo (Humans), Simia (monkeys and thirds), Lemur (lemurs and lorises) and Vespertilio. The main characteristics of the order lie in the upper front teeth and two mammary glands on their pectorals [63]. For studies of their genetic variability, population genetic analyses are available for the species Alouatta caraya of Argentina, Paraguay and Brazil [30]. The situation is different for the order Rodentia, considered the most diverse among mammals, characterised by a gnawing dentition composed of a pair of lower incisors with continuous growth [64]. Studies on genetic variability have prioritised their high variation due to environmental homogeneity, in addition to very little inbreeding in their populations [65]. For these orders,

Table 4. Main species studied and their heterogeneity values in the orders Proboscidea, Primates and Rodentia.

Species	Main Threats	Conservation Strategy	Genetic Sampling	Expected and Observed Heterocigosity	Reference
Elephas maximus	Habitat fragmentation and reduced gene flow	Semi-captive species	Blood	He = 0.64 Ho = 0.63	[66]
Elephas maximus	Illegal hunting	Semi-captive species	Faeces	He = 0.67 Ho = 0.57	[29]
Alouatta caraya	Deforestation, agriculture and livestock	Reintroduction and management units	Faeces and tissue	He = 0.44 Ho = 0.46	[30]
Sciurus griseus *	Habitat loss and fragmentation due to urbanization	-	Hair	He = 0.49 Ho = 0.45	[31]

* Average heterozygosity across populations, because more than two populations are involved.

describes the species of each order, as well as their heterozygosity values, problems and conservation strategies in each of the populations studied from a genetic point of view.

6.2. Population Balance and Inbreeding Coefficients

For the order Proboscidea, population variability was high in both studies, with expected heterozygosis (He) values ranging from 0.21 to 0.90 according to the Hardy–Weinberg equilibrium hypothesis in Elephas maximus specimens [29]. However, these investigations did not consider the calculation of inbreeding coefficients. In the order Primates for the species Alouatta caraya, variability was high with a negative inbreeding coefficient (Fis = −0.06), but there was no population imbalance [30]. In the case of rodents of Sciurus griseus, the deviation was significant but the variability was moderate and there was a value of Fis = −0.018 [31]. The orders analyzed in this study are sparser and are different from the others mentioned above. However, small mammals are not so affected and their population variability and viability is compromising for their subsistence. It is worth mentioning that the estimation of inbreeding coefficients can help all three orders, understanding that it is a useful technique for any type of mammal. The samples analyzed were mostly hair samples. However, few studies are considered for the very large orders.

7. 2015–2025: Methodological Shifts and Best Practices in Non-Invasive Population Genomics

In 2015–2018, reduced-representation methods (RAD and ddRAD) were piloted on non-invasive DNA but suffered from allele dropout and locus dropout due to fragmentation and inhibitors. From ~2018 onward, two strategies rose to prominence for degraded templates: amplicon panels (GT-seq)—hundreds of short (<120–150 bp) targets multiplexed per sample and hybridisation capture of short SNP-bearing fragments using custom baits, often with single- or unique-molecular-index (UMI) libraries to track duplicates. These approaches improved locus recovery and cross-study comparability while keeping per-sample costs manageable. STRs remain useful for continuity with legacy datasets and for certain kinship applications, but SNP panels now dominate where non-invasive sampling is routine.

Early non-invasive studies emphasised detection, minimum population size, or broad diversity (He) and FIS estimates, constrained by 8–20 STR loci and genotyping error. With hundreds–thousands of SNPs, 2019–2025 studies routinely estimated relatedness, assignment, hybridisation, and effective population size with tighter precision, and detected fine-scale structure and sex-biased dispersal. Demographic inference shifted from methods requiring long contiguous haplotypes (unsuitable for fragmented DNA) to site–frequency–spectrum and genotype likelihood approaches that tolerate low coverage. Kinship and assignment accuracy improved notably as panels incorporated ancestry-informative and immune-linked SNPs while controlling for missingness.

Best practice converged on (1) multi-tube replication at extraction and PCR to quantify allelic dropout; (2) small-amplicon design (<120 bp) or capture of short inserts; (3) blocking oligos or predator-specific PNA and LNA clamps to reduce non-target amplification in faecal samples; (4) UDG treatment when deamination patterns suggest ancient-DNA-like damage; (5) consistent use of UMIs and dual indices to control index hopping; and (6) rigorous negative controls and extraction blanks. Field preservatives shifted from variable ethanol/silica protocols to validated buffers (e.g., Longmire’s or equivalent commercial stabilisers) with documented inhibitor removal at extraction.

The period saw a transition from hard genotype calls to probabilistic genotype-likelihood frameworks that explicitly model dropout and depth (e.g., ANGSD-like pipelines), followed by error-aware filtering (allele balance, strand bias, depth bounds) and replication-informed consensus. For amplicon and capture panels, per-locus performance metrics (on-target rate, duplicate fraction, allelic dropout rate) are now reported alongside population parameters, improving reproducibility and meta-analysis.

While shotgun sequencing remains inefficient for low endogenous DNA, target capture reduced sequencing waste and made SNP-scale inference feasible for programs with modest budgets. Amplicon panels require upfront design but enable scalable monitoring. Method selection increasingly reflects decision frameworks balancing aims (assignment vs. N_e), sample type (hair vs. faeces), DNA quality, and resources.

Terrestrial eDNA and metabarcoding matured for detection and community composition (soil, leeches, carrion flies), but only rarely support population-genomic inference at present. Our synthesis positions eDNA as complementary for occurrence and surveillance, while individual-assignable non-invasive samples (scat/hair/urine) remain the backbone for population genetics.

The last decade normalised explicit reporting of permits, animal welfare minimisation, and FAIR data deposition of both raw reads and panel definitions (probe/amplicon lists), enabling re-use and cumulative gains.

Across 2015–2025, non-invasive applications (including “both”) increased steadily and consistently outnumbered invasive-only studies in recent years, reflecting the shift toward minimally disruptive sampling at scale. Marker usage was dominated by STR and mtDNA early in the period, with an evident rise of SNP/NGS-oriented approaches toward the latter years, consistent with the review’s narrative on the transition from small STR panels to SNP-based genotyping suitable for degraded templates. Taken together, the timeline highlights the field’s methodological drift toward ethical, logistically scalable sampling coupled with higher-resolution markers, aligning population-genetic inference with contemporary conservation needs (Figure 3).

8. Future Orientations

Future research should focus on improving the quality and quantity of DNA obtained by non-invasive methods, optimizing specific molecular protocols for each type of biological sample. Further comparative studies between different taxonomic orders are also needed to generalize results and recommendations. It is recommended to explore new molecular techniques such as massive sequencing and advanced bioinformatics methods that can compensate for the limitations on the quality of DNA obtained from non-invasive methods. Further research on the ethical impact and social perceptions related to the use of invasive methods in different cultural and social contexts would also be beneficial.

9. Limitations of the Study

An important limitation of this review was the restriction to studies published in English only, which could exclude relevant research published in other languages. In addition, the methodological heterogeneity observed among the reviewed studies limited direct comparison of results, making it difficult to generalize definitive conclusions. Another limitation is the absence of specific analysis on the direct environmental impact associated with each sampling method, which could be relevant for future research.

10. Conclusions

This systematic review collated data from 31 peer-reviewed studies, published between 2015 and April 2025, which compared the effectiveness of invasive versus non-invasive genetic sampling techniques in wild terrestrial mammals. While invasive sampling consistently yielded the highest DNA integrity and facilitated genomic analyses, it necessitated the capture and anaesthesia of animals, thereby increasing costs and risks to animal welfare. Conversely, non-invasive methodologies reduced disruption and facilitated population-scale coverage; nevertheless, they were plagued by issues including diminished DNA quality, allele loss, and contamination. The concept of environmental DNA (eDNA) represents a novel methodology that facilitates the estimation of presence/absence and diversity without the necessity for direct contact. Conversely, recent advances in conservation techniques (e.g., silica desiccation, cold chain, chelating buffers) now enable the reliable amplification of DNA from degraded sources. Nevertheless, the phenomenon of marker bias remains in evidence. Despite the existence of SNP and CNV platforms, STR predominate. The statistical robustness of these methods is contingent upon the implementation of corrections for allele loss and Hardy–Weinberg deviations. Studies that employed multitube consensus attained reduced genotyping error rates.