J. CETACEAN RES. MANAGE. 18: 33–80, 2018 33 Guidelines for genetic data analysis ROBIN S. WAPLES1, A. RUS HOELZEL2, OSCAR GAGGIOTTI3, RALPH TIEDEMANN4, PER J. PALSBØLL5, FRANK CIPRIANO6, JENNIFER JACKSON7, JOHN W. BICKHAM8 AND AIMÉE R. LANG9 Contact e-mail: [email protected] ABSTRACT The IWC Scientific Committee recently adopted guidelines for quality control of DNA data. Once data have been collected, the next step is to analyse the data and make inferences that are useful for addressing practical problems in conservation and management of cetaceans. This is a complex exercise, as numerous analyses are possible and users have a wide range of choices of software programs for implementing the analyses. This paper reviews the underlying issues, illustrates application of different types of genetic data analysis to two complex management problems (involving common minke whales and humpback whales), and concludes with a number of recommendations for best practices in the analysis of population genetic data. An extensive Appendix provides a detailed review and critique of most types of analyses that are used with population genetic data for cetaceans. KEYWORDS: ABUNDANCE ESTIMATE; BREEDING GROUNDS; CONSERVATION; DNA FINGERPRINTING; FEEDING GROUNDS; GENETICS; HUMPBACK WHALE; MIGRATION; MINKE WHALE; REPRODUCTION; TAXONOMY INTRODUCTION before the analyses considered here begin, the DNA data Recently, guidelines were adopted for quality control of quality-control guidelines have been consulted and followed DNA data intended for use within the International Whaling to the extent possible, and that any substantial deviations Commission (IWC, 2009; 2015a). Once the data have have been documented and explained. been collected, the next step is to analyse the data and make As discussed in detail later, genetic information can inferences that are useful for addressing practical problems provide insights relevant to many types of problems in the management of cetaceans. This is a complex exercise associated with conservation and management of living for two major reasons: (1) many methods can be used natural resources. Among other applications, genetic data can to analyse genetic data, and an equally wide range of be used to: computer software is available to conduct data analyses; (1) identify and delimit biological species, subspecies and and (2) a key objective is to inform those involved in populations; cetacean management who do not have a background in population genetics. For these reasons, it has been suggested (2) provide or improve estimates of census population size that it would be useful to have a document that provides (N) and effective population size (Ne); guidelines for the analysis of population genetic data for use (3) help track contemporary movements of individuals, as in a management context. Although it is not possible (nor well as estimate long-term levels of connectivity among is it desirable) to prescribe specific procedures for all populations; analyses of population genetic data, it can help to provide general guidelines for some of the more common types of (4) quantify genetic diversity within populations and provide analyses conducted in a management context. The latter is insights into past bottlenecks and population expansions; the objective of this paper. Emphasis is on a general discussion of issues involved in genetic data analysis (5) help resolve mixtures of individuals originating from rather than detailed comments about specific computer different breeding populations; and software, but some popular programs will be discussed to (6) track products through the marketplace. emphasise particular points. Given the many analytical methods (and software packages) available, to focus on However, the most widespread practical application, those most relevant to a particular study, the discussion particularly in the IWC context, is for the study of stock has been organised around some common management structure (genetic differentiation among populations). Before problems one might try to address with genetic data. discussing details of particular genetic analyses, some of the These problems are identified below. It is assumed that key issues involved with assessing stock structure are 1 NOAA Fisheries, Northwest Fisheries Science Center, Seattle WA 98112, USA. 2 Department of Biosciences, Durham University, Durham DH1 3LE, UK. 3 Scottish Oceans Institute, East Sands, University of St. Andrews, KY16 8LB, UK. 4 Institute of Biochemistry and Biology, University of Potsdam, D-14476 Potsdam, Germany. 5 Groningen Institute for Evolutionary Life Sciences, University of Groningen, The Netherlands. 6 California Academy of Sciences, San Francisco CA 94118, USA. 7 British Antarctic Survey, Cambridge CB3 0ET, UK. 8 Texas A&M University, College Station, TX 77843. 9 NOAA Fisheries, Southwest Fisheries Science Center, La Jolla CA 92037, USA. 34 WAPLES et al.: GUIDELINES FOR GENETIC DATA ANALYSIS summarised and best practice for assessing issues that may found within and among oceans (e.g. bottlenose dolphins, be complex or challenging are highlighted. Natoli et al., 2005). Common questions that arise in conservation and Two major issues related to stock structure/population management of living natural resources include the differentiation following: Is the differentiation among groups of individuals Two major issues that arise in applying genetic data to strong enough that they should be considered separate problems in stock structure are: (1) identification of threshold populations or stocks? Is any statistically significant (e.g. levels of population differentiation that require separate- P < 0.05) departure from panmixia sufficient to warrant stock management to achieve stated objectives; and (2) using recognition as separate stocks? If not, how strong must the genetic and other data to determine whether the system under differentiation be? Unfortunately, there are no generally consideration is above or below this threshold. applicable answers to these questions, since the relevant degree of population differentiation depends on the Identifying threshold levels of population differentiation conservation/management objectives, the risks associated Population differentiation occurs along a continuum (Fig. 1). with adopting different management strategies, and society’s At one extreme (completely random mating), every tolerance of the resulting consequences. Some general individual has an equal probability of mating with any other management objectives/considerations include: individual. This situation is referred to as panmixia; although this is an idealised scenario not known to exist in any natural (1) ‘Management units’ or ‘populations’ or ‘stocks’ must be population, panmixia is typically adopted as the null considered separately because of a legal mandate. In the hypothesis against which to compare alternative hypotheses US, federal laws that include this type of mandate include that involve various degrees of departure from random the Endangered Species Act, the Marine Mammal mating. The other extreme of the population-differentiation Protection Act, and the Magnuson-Stevens Fishery continuum is characterised by complete isolation among Conservation and Management Act (see Waples et al. 2008 locally panmictic groups of individuals. In nature, populations for discussion). Canada has similar provisions in its federal in general are neither completely panmictic nor completely Species at Risk Act (2002), as do the Biodiversity Law of isolated; instead, they typically are characterised by Costa Rica (1992), Australia’s Endangered Species intermediate levels of differentiation and linked by restricted Protection Act (2002), and South Africa’s National but non-zero levels of migration. In addition, the degree of Environmental Management Biodiversity Act (2004). connectivity among populations often changes over time. (2) Sustainable harvest for management stocks should be Several types of data, including genetic information, can maximised while preventing/minimising impacts on help to determine where a particular species falls on the stocks that cannot withstand harvest, because: population-differentiation continuum depicted in Fig. 1. In some cetacean species, the level of genetic differentiation (a) locally depleted stocks might take a long time to among geographic areas within major ocean basins is rebuild, and/or relatively low (toward the panmictic end of the continuum), (b) local extirpation might represent an irreversible loss whereas evidence for higher population differentiation is of biodiversity. often found among populations from different oceans (e.g. sperm whales; see Alexander et al., 2016). In other cases, Although such general considerations are useful for unexpectedly high levels of population differentiation are providing context, they are qualitative rather than Fig. 1. The continuum of population differentiation. Each circle represents a group of individuals that might or might not be a separate population or stock. Four generic scenarios, with varying degrees of connectivity (geographical overlap and/or migration), are identified along the continuum: (A) Complete independence. (B) Modest connectivity. (C) Substantial connectivity. (D) Panmixia (circles are completely congruent). Reproduced from Waples and Gaggiotti (2006). J. CETACEAN RES. MANAGE. 18: 33–80, 2018 35 quantitative and by themselves will not produce repeatable outcomes. That is, you could not provide guidance this general to independent