- Email: [email protected]
What’s missing from avian global diversification analyses?
Accepted Manuscript Review What’s missing from avian global diversification analyses? Sushma Reddy PII: DOI: Reference:
S1055-7903(14)00151-1 http://dx.doi.org/10.1016/j.ympev.2014.04.023 YMPEV 4888
To appear in:
Molecular Phylogenetics and Evolution
Received Date: Revised Date: Accepted Date:
1 December 2013 18 April 2014 18 April 2014
Please cite this article as: Reddy, S., What’s missing from avian global diversification analyses?, Molecular Phylogenetics and Evolution (2014), doi: http://dx.doi.org/10.1016/j.ympev.2014.04.023
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
-1What’s missing from avian global diversification analyses? Sushma Reddy Biology Department, Loyola University Chicago, Chicago, IL 60660, USA; [email protected]
-2Abstract The accumulation of vast numbers of molecular phylogenetic studies has contributed to huge knowledge gains in the evolutionary history of birds. This permits subsequent analyses of avian diversity, such as how and why diversification varies across the globe and among taxonomic groups. However, available genetic data for these meta-analyses are unevenly distributed across different geographic regions and taxonomic groups. To comprehend the impact of this variation on the interpretation of global diversity patterns, I examined the availability of genetic data for possible biases in geographic and taxonomic sampling of birds. I identified three main disparities of sampling that are geographically associated with latitude (temperate, tropical), hemispheres (East, West), and range size. Tropical regions, which host the vast majority of species, are substantially less studied. Moreover, Eastern regions, such as the Old World Tropics and Australasia, stand out as being disproportionately undersampled, with up to half of communities not being represented in recent studies. In terms of taxonomic discrepancies, a majority of genetically undersampled clades are exclusively found in tropical regions. My analysis identifies several disparities in the key regions of interest of global diversity analyses. Differential sampling can have considerable impacts on these global comparisons and call into question recent interpretations of latitudinal or hemispheric differences of diversification rates. Moreover, this review pinpoints understudied regions whose biota are in critical need of modern systematic analyses.
-3Keywords: Biodiversity; Birds; Diversity Patterns; DNA sequences; Sampling bias; Taxonomy
1.1 Introduction There is a critical need to understand global patterns of biodiversity to increase knowledge of ecology and evolutionary history and to predict the effects of anthropogenic impacts (Gaston, 2000; Hurlbert and Jetz, 2007; Kier et al., 2009; Myers et al., 2000). A major goal of global biodiversity studies is to examine how and why diversity varies across the world. These studies range from describing diversity patterns such as taxonomic richness and endemism to investigating why some regions might produce and/or accumulate species at different rates (Hawkins et al., 2007; Jetz and Fine, 2012; Jetz et al., 2012; Mccain, 2009; Orme et al., 2006). Examining diversification by estimating speciation and extinction rates can help to understand why diversity varies across different landscapes. For instance, a longstanding question is why the tropics are more diverse than temperate regions (Jablonski et al., 2006; Weir and Schluter, 2007; Willig et al., 2003) and much of the debate has been about whether this is due to the tropics acting as cradles (producing many new species) or museums (supporting the accumulation of many species). Examining and testing hypotheses of global patterns relies on having sufficient comparative data. Birds, with about 10,000 described species (Clements et al., 2012; Dickinson et al., 2003), are generally well-studied using modern systematic tools (i.e. with genetic, morphological, and/or distributional
-4data). The accumulation of genetic data, with roughly two-thirds of all described avian species having some published DNA sequence data, has prompted many recent studies to examine large-scale patterns of avian diversification to test hypotheses such as the latitudinal gradient (Hawkins et al., 2007; Jetz et al., 2012; Lijtmaer et al., 2011; Ricklefs, 2006; Weir and Schluter, 2007), East-West hemispheric differences (Jetz et al., 2012; Lijtmaer et al., 2011), densitydependent diversification rates (Morlon et al., 2010; Phillimore and Price, 2008), and ecological correlates of diversity (Orme et al., 2006; Thomas et al., 2008). Furthermore, the vast accumulation of genetic data enabled a recent study to build a 'complete' phylogeny of all living birds (Jetz et al., 2012), filling in missing species by constraining priors using taxonomy. This allowed the study to analyze diversification rates by representing all extant lineages but implicitly assumes that the contribution of evolutionary rates from missing species is minimal. Jetz et al. (2012) used their phylogeny to study patterns of avian diversification across geographic space and evolutionary time. One major conclusion drawn by the authors was that the disparity in diversification rates was not so profound when comparing temperate and tropical regions, but rather that there was a significant difference in rates between the Eastern and Western Hemispheres or Old and New Worlds (Jetz et al., 2012). Furthermore, they allude to an intriguing pattern showing that "avian assemblages in Australia, Southeast Asia, Africa, and Madagascar are characterized by particularly low average rates" of diversification compared to the global mean (Jetz et al., 2012, p. 445). Here, I examine if there is a bias in terms of availability of genetic data from
-5different regions that might mislead such global analyses. Using taxonomy as a proxy for phylogenetic relationships is not an appropriate assumption for many reasons. Taxonomic classifications in birds have been demonstrated to be exceedingly inaccurate in almost every avian molecular systematics study. Avian taxonomic groups at all levels – orders, suborders, families, subfamilies, genera, species, and subspecies – are not stable (e.g., Alström et al., 2006; Barker et al., 2004; Beresford et al., 2005; Cibois and Cracraft, 2004; Hackett et al., 2008; Oliveros et al., 2012; Olsson et al., 2005; Reddy and Cracraft, 2007). Extreme examples of misclassification include species classified based on phenotypic similarities that are now shown to be in highly divergent lineages, in some cases on the order of families (Alström et al., 2014; Cibois, 2003; Cibois et al., 2001; Reddy and Cracraft, 2007). Moreover, an increasing number of studies point out that even shallow evolutionary divergences are complicated with numerous genera and species not being monophyletic (Burns et al., 2014; Lovette et al., 2010; Moyle et al., 2012; Olsson et al., 2005) or recognized (Siler et al., 2014; Tobias et al., 2008). Calculations of diversification rates can be significantly impacted when these lineages, both deeply divergent and shallow branches, go unrecognized. While diversification analyses such as the Jetz et al. (2012) paper are useful and will no doubt lead to subsequent studies, they must be interpreted in light of how well biodiversity is sampled across the globe. This study is not intended to be a critique of these studies or their methods. Rather, I am interested in using available genetic data to examine how much we know about
-6avian species from different regions of the world. Differential knowledge of regional biodiversity, which is impacted by the feasibility/logistics of acquiring samples and scientific interest, could influence the resulting patterns of global analyses. In all of these studies, the implicit assumption is that missing species will not contribute any different information from that derived from available data. Yet this is true only if missing species are randomly distributed, minimal, or contribute the same signal as species with data. Although a few methods have been proposed to accommodate incomplete sampling for estimating diversification rates (Brock et al., 2011; Pybus and Harvey, 2000; Reddy et al., 2012), simulation studies show that results can be unreliable when clade sampling falls below 75-80% completeness (Brock et al., 2011; Cusimano and Renner, 2010) and varied whether or not missing taxa were randomly distributed. The effect of missing species in the avian phylogeny can only be truly tested when they are, in time, sequenced and included in these comparative studies. Examination of possible biases in sampling is key to understanding the limitations of the results and interpretation of diversification studies. Our knowledge of biodiversity patterns can be biased by both geographic and taxonomic sampling (Boakes et al., 2010; Brock et al., 2011; Cusimano and Renner, 2010; Reddy and Dávalos, 2003; Tobias et al., 2008). Here, I use genetic data as an indicator of whether a species has been examined in a recent systematic study to assess if differences in research effort across the globe are substantial enough to impact studies of diversification patterns. I examined the current state of knowledge of bird phylogeny for sampling biases in terms of
-7geography and taxonomy. Birds are generally better studied that other taxonomic groups, therefore it is likely that the patterns uncovered here are similar or more pronounced in less-studied groups.
1.2 Methods I used the genetic data compiled by Jetz et al., (2012) as indicative of available nucleotide sequences for birds (as of April 2011). Although several recent large-scale phylogenies have been published since, (e.g., Barker et al., 2013; Derryberry et al., 2011; Moyle et al., 2012), my assumption here is that the proportional geographic representation is still similar. Jetz et al. recognized 9,993 species, of which 6,663 had some published genetic data. I categorized all bird species into binary distinctions of with (WiG) or without (NoG) genetic data. Unfortunately, there is no geographic data associated with Genbank accessions, a limitation that still needs to be addressed (Marques et al., 2013). Since linking geographic coordinates to each sample was not possible, I made the simple assumption that a species could have been sampled anywhere across its geographic range. For this analysis, I did not distinguish how much data were available for each taxon, but rather considered species as WiG if they had at least one sequence published. While this might bias the analysis in some ways, for instance species with only one sequence are treated the same as other species for which there are multiple sequences, a more precise analysis is unfortunately not feasible at present given the limitations of Genbank deposits. The Jetz et al. study focused on sequences from 15 commonly used loci. While
-8some species might be represented in GenBank with sequences from other loci only, these are a small minority and would not be useful in a comparative context of a larger phylogeny.
1.2.1 Geographic sampling To examine geographic biases in sampling, I used a dataset of species distributions of all birds assembled and distributed through Birdlife International and NatureServe (BirdLife International, Inc. and NatureServe, 2012) to geographically plot the distribution of WiG and NoG species. In ArcGIS (v.10; Environmental Systems Research Institute, Redlands, CA.), I used the native, resident or breeding distributions and combined shapefiles of species with disjunct distributions (using the ‘Dissolve’ function). I then created a 5° by 5° grid across globe (‘Create Fishnet’ function) and counted the total WiG and NoG species in each grid cell (‘Spatial Join’ and ‘Intersect’ functions). Grid cells of different sizes and projections should still yield similar results since the main comparison of the analysis is the proportion of missing species per area. I focused on grid cells that incorporated terrestrial regions by excluding cells with <30 total species, a reasonable assumption because even remote islands usually host substantially more species. I further examined the relationship of range size to genetic sampling to test if NoG species are more likely to have small ranges and therefore were presumably harder to sample. Taking the total area of each species’ resident or breeding range polygon, I compared the frequency distribution of range sizes for
-9WiG and NoG species. I used an unpaired t-test with Welch’s correction and F test to statistically assess if the means and variances of the two distributions were significantly different using Prism (v. 5; GraphPad Software, San Diego, CA).
1.2.2 Taxonomic sampling To examine taxonomic skewness of genetic data, I used the 129 crown clades in the backbone of the Jetz phylogeny to examine if less sampled groups tended to be from certain geographic regions. I also examined whether there was a relationship between clade size, genetic completeness (i.e. percent of WiG taxa), and average data decisiveness (DD; see Jetz et al., 2012), which might skew the results. For each pairwise comparison, I calculated the correlation coefficient in the R statistical package. While some of these potential issues were inspected by Jetz et al., I examined these correlations to further test the impact of missing species. A negative correlation between clade size and completeness would indicate that complete sampling is harder for larger clades. Data decisiveness, a measure of the ability of incomplete data matrices to resolve trees (Sanderson et al., 2010), was measured as the proportion of branches that the data could resolve across 1000 random trees (Jetz et al. 2012). A relationship between DD and completeness indicates that missing data impacts the determination of phylogenetic relationships and an relationship between DD and clade size could indicate that the number of taxa influences how branches are resolved.
- 10 -
1.3 Results Figure 1 shows maps of the geographic distribution of total species richness, number of NoG species, and proportion of NoG species (number of NoG species divided by total species) in each 5x5 degree grid cell. Global patterns of species richness are similar to those discussed previously (Orme et al., 2005), with highest richness being in the tropical montane regions (i.e. Andes, Eastern Africa, Himalayas). Areas with most NoG species coincide with areas of highest richness. However, when examining the proportion of NoG species, the pattern is remarkably different. Areas with proportionally higher missing data—up to 50% NoG species—are concentrated in the Old World tropics and Australasia. Species range sizes of birds span several orders of magnitude. Figure 2 shows the distribution of range sizes for NoG and WiG species. While they both extend over the same breadth, the means and variances of these categories are significantly different. A higher percentage of NoG species have smaller ranges than WiG species. Figure 3 shows the relationship between completeness, clade size, and data decisiveness. The 129 major clades ranged in size from 5 to 460 species. Within these groups, completeness or percentage of WiG species varied from 21.6% to 100%. Data decisiveness, a measure of the impact of missing data on recovering relationships calculated as the proportion of distinguished branches across 1000 trees (Jetz et al., 2012; Sanderson et al., 2010), spanned from 0.23
- 11 to 1 in these clades. There was no correlation between clade size and average decisiveness (correlation coefficient=0.025), a weak negative correlation between clade size and completeness (-0.229), and a moderate correlation between completeness and average decisiveness (0.402). Table 1 shows the clades at the low and high ends of completeness. Table 1a shows clades with 50% or less completeness. These 23 clades (or 17% of all clades) encompass more than 36% of the total number of NoG species. Of these, 9 clades are exclusively distributed in the Old World, 4 are only in the New World, 1 in Australasia, and 9 are widespread. Furthermore, 16 of the 23 clades predominately occur in tropical regions, while the others generally occur in both tropical and temperate zones. These clades were about evenly spread between the distinctions of passerines and non-passerines. Table 1b shows clades with more than 90% completeness. These are 27 clades that are generally smaller (20 clades are less than or equal to 20 species) with a majority being nonpasserines. These clades are almost evenly distributed across all regions.
1.4 Discussion The increasing number of phylogenetic studies has made a near-complete avian tree of life tenable. Yet efforts of modern biodiversity studies, as measured by genetic analyses, show substantial differences across the globe. In this study, I have identified three main disparities associated with genetic sampling of birds across the world, including between high and low latitudes, among Eastern and Western hemispheres, and across small and large ranges. Given the drastic
- 12 variation in sampling, interpretations of meta-analyses have to be taken in light of the limitations of the data used therein. Many of the disparities in sampling effort have crucial implications for the regions of interest in most analyses of global diversity patterns. Not surprisingly, regions with highest species diversity also have the most NoG species. When looking at both the total number of species as well as the proportion of the community that is NoG, there is substantial disparity across latitudes. Temperate regions, especially North America and Europe, comprise the most studied avifaunas despite having lower species richness. The tropics are poorly sampled in genetic studies. These areas are highlighted as having overall high numbers of NoG species (Fig. 1) and, additionally, the majority of clades with less than 50% completeness are found primarily in the tropics (Table 1). Within the tropics, there is a clear discrepancy between sampling in the Western and Eastern Hemispheres. The New World is better studied than other tropical regions. Some parts of the Old World and Australasia are so poorly known that up to half the communities in these regions are missing from recent studies. These are exactly the regions in the Jetz et al. analysis that showed much lower average diversification rates than the global mean, calling into question whether these calculations are biased by uneven sampling. The Old World and Australasia, in contrast to the Neotropics, are logistically more heterogeneous and have fewer researchers focused on these regions, especially in modern studies. The analysis of NoG and WiG species range sizes implies a connection
- 13 between range size and accessibility. The distribution of WiG species is skewed towards wider ranges, while the NoG species tend to have smaller range sizes. However, the extents of both distributions are similar, indicating there are other factors that impact how biotic surveys and studies are conducted. There are WiG species with extremely restricted ranges and likewise NoG species with very wide ranges (mainly seabirds). Estimates of diversification rates of incompletely sampled clades are known to be unreliable (Brock et al., 2011; Cusimano and Renner, 2010; Pybus and Harvey, 2000). Simulation studies have shown that error rates for diversification estimates increase substantially when clade completeness is less than 75-80% (Brock et al., 2011; Cusimano and Renner, 2010). This would include 80 of the 129 major clades in the Jetz et al. analysis. Although Jetz et al. (2012) compensated for missing species by placing NoG taxa onto the phylogeny by using taxonomy as a proxy for evolutionary history, this still resulted in rates estimates being essentially based on WiG taxa (see their supplementary discussion figure 2). The underlying result is that NoG species do not contribute additional information for diversification rates than those derived from the prior constraints in these Bayesian analyses. Furthermore, my analysis shows genetic sampling completeness is correlated to data decisiveness for phylogenetic relationships within clades (Fig. 3). When taxa are missing phylogenetic information, studies often substitute taxonomy as an indicator of evolutionary history. However, many studies (ironically, using genetic data) have shown that traditional groupings and
- 14 relationships implied by taxonomy are grossly erroneous. These include incorrect assessments of species-limits, monophyly, and phylogenetic relationships, which have further obscured our understanding of biodiversity patterns (for examples see Alström et al., 2011; Barker et al., 2004; Beresford et al., 2005; Fuchs et al., 2006; Hackett et al., 2008; Moyle et al., 2012; Oliveros et al., 2012; Reddy et al., 2012). This underscores the need for increased biodiversity research in understudied regions to clarify these inaccuracies. Additionally, continued surveys of little known regions are also key to identifying new, overlooked or cryptic species (see Athreya, 2006; Voelker et al., 2010).
1.5 Conclusions Differential knowledge across regions impacts interpretations of global biogeographic patterns and has critical implications for the informative power of large-scale meta-analyses. This study shows that there are considerable discrepancies in terms of which areas are well-studied using modern systematic analyses. Many of the under-sampled regions are key areas of focus when addressing big question in global biodiversity. The early biogeographic history of many groups of birds is still debated and some of the understudied regions, such as Asia and Australasia, have been proposed to play a key role in the diversification of some groups (Barker et al., 2004; Cracraft, 2001; Jønsson et al., 2011). There is an urgent need to prioritize studies in these poorly known regions to better understand the evolution of modern birds, both in terms of enumerating true diversity and examining their history of diversification. This is especially very
- 15 important in the most severely understudied regions of the Old World Tropics and Australasia, where areas of high diversity coincide with those most at risk by anthropogenic pressures (Hoffmann et al., 2010; Orme et al., 2005). Addressing these major gaps in our knowledge of avian diversity will require intense efforts to target analyses of these regions. Nevertheless, some of these efforts might be hindered by access to fresh tissue samples or political and cultural impediments to biodiversity surveys. While I am not advocating that genetic studies are the only way to conduct modern systematic analyses, these data do contribute a significant portion towards our overall understanding of diversity. If these key geographic areas continue to be holes in our map of life, we will miss out on a truly comprehensive view of biodiversity.
Acknowledgements This work was conducted with financial support from the US National Science Foundation (DEB-0962078) and facilitated by the National Centre for Biological Sciences (India). Access to species distribution data was generously provided by BirdLife International and NatureServe. I am grateful to D. Treering and V. Varma for advice regarding GIS analyses and J. Bates, L. Davalos, P. Makovicky, R. Nandini, H. Skeen, U. Ramakrishnan, V.V. Robin, and two anonymous reviewers for constructive comments to improve this manuscript.
References Alström, P., Ericson, P.G.P., Olsson, U., Sundberg, P., 2006. Phylogeny and classification of the avian superfamily Sylvioidea. Molecular Phylogenetics
- 16 and Evolution 38, 381–397. Alström, P., Fregin, S., Norman, J.A., Ericson, P.G.P., Christidis, L., Olsson, U., 2011. Multilocus analysis of a taxonomically densely sampled dataset reveal extensive non-monophyly in the avian family Locustellidae. Molecular Phylogenetics and Evolution 58, 513–526. Alström, P., Hooper, D.M., Liu, Y., Olsson, U., Mohan, D., Gelang, M., Le Manh, H., Zhao, J., Lei, F., Price, T.D., 2014. Discovery of a relict lineage and monotypic family of passerine birds. Biology Letters 10, 20131067. Athreya, R., 2006. A new species of Liocichla (Aves: Timaliidae) from Eaglenest Wildlife Sanctuary, Arunachal Pradesh, India. Indian Birds 2, 82–94. Barker, F.K., Burns, K.J., Klicka, J., Lanyon, S.M., Lovette, I.J., 2013. Going to Extremes: Contrasting Rates of Diversification in a Recent Radiation of New World Passerine Birds. Systematic Biology 62, 298–320. Barker, F.K., Cibois, A., Schikler, P., Feinstein, J., Cracraft, J., 2004. Phylogeny and diversification of the largest avian radiation. Proc Natl Acad Sci USA 101, 11040–11045. Beresford, P., Barker, F.K., Ryan, P., Crowe, T.M., 2005. African endemics span the tree of songbirds (Passeri): molecular systematics of several evolutionary “enigmas.” Proceedings of the Royal Society, B 272, 849–858. BirdLife International, Inc., NatureServe, 2012. Bird Species Distribution Maps of the World. Version 2.0. Boakes, E.H., McGowan, P.J.K., Fuller, R.A., Chang-qing, D., Clark, N.E., O'Connor, K., Mace, G.M., 2010. Distorted views of biodiversity: spatial and temporal bias in species occurrence data. PLoS Biology 8, e1000385. Brock, C.D., Harmon, L.J., Alfaro, M.E., 2011. Testing for temporal variation in diversification rates when sampling is incomplete and nonrandom. Systematic Biology 60, 410–419. Burns, K.J., Shultz, A.J., Title, P.O., Mason, N.A., Barker, F.K., Klicka, J., Lanyon, S.M., Lovette, I.J., 2014. Molecular Phylogenetics and Evolution. Molecular Phylogenetics and Evolution 75, 41–77. Clements, J.F., Schulenberg, T.S., Iliff, M.J., Sullivan, B.L., Wood, C.L., 2012. The eBird/Clements checklist of birds of the world: Version 6.7. Downloaded from http://www.birds.cornell.edu/clementschecklist/downloadable-clementschecklist. Cibois, A., 2003. Mitochondrial DNA phylogeny of babblers (Timaliidae). The Auk 120, 35–54. Cibois, A., Cracraft, J., 2004. Assessing the passerine “Tapestry”: phylogenetic relationships of the Muscicapoidea inferred from nuclear DNA sequences. Molecular Phylogenetics and Evolution 32, 264–273. Cibois, A., Slikas, B., Schulenberg, T., Pasquet, E., 2001. An endemic radiation of Malagasy songbirds is revealed by mitochondrial DNA sequence data. Evolution 55, 1198–1206. Cracraft, J., 2001. Avian evolution, Gondwana biogeography and the Cretaceous-Tertiary mass extinction event. Proceedings of the Royal Society, B 268, 459–469. Cusimano, N., Renner, S.S., 2010. Slowdowns in Diversification Rates from Real
- 17 Phylogenies May Not be Real. Systematic Biology 59, 458–464. Derryberry, E.P., Claramunt, S., Derryberry, G., Chesser, R.T., Cracraft, J., Aleixo, A., Pérez-Emán, J., Remsen, J.V., Jr, Brumfield, R.T., 2011. Lineage Diversification and Morphological Evolution in a Large-Scale Continental Radiation: the Neotropical Ovenbirds and Woodcreepers (Aves: Furnariidae). Evolution 65, 2973–2986. Dickinson, E.C., Bahr, N., Dowsett, R., Pearson, D., Remsen, V., Roselaar, C.S., Schodde, D., 2003. The Howard and Moore Complete Checklist of Birds of the World, revised and enlarged. 3rd edition. Christopher Helm, London, UK. Fuchs, J., Fjeldså, J., Bowie, R.C.K., Voelker, G., Pasquet, E., 2006. The African warbler genus Hyliota as a lost lineage in the Oscine songbird tree: molecular support for an African origin of the Passerida. Molecular Phylogenetics and Evolution 39, 186–197. Gaston, K., 2000. Global patterns in biodiversity. Nature 405, 220–227. Hackett, S.J., Kimball, R.T., Reddy, S., Bowie, R.C.K., Braun, E.L., Braun, M.J., Chojnowski, J.L., Cox, W.A., Han, K.-L., Harshman, J., Huddleston, C.J., Marks, B.D., Miglia, K.J., Moore, W.S., Sheldon, F.H., Steadman, D.W., Witt, C.C., Yuri, T., 2008. A phylogenomic study of birds reveals their evolutionary history. Science 320, 1763–1768. Hawkins, B.A., Diniz-Filho, J.A.F., Jaramillo, C.A., Soeller, S.A., 2007. Climate, Niche Conservatism, and the Global Bird Diversity Gradient. American Naturalist 170, S16–S27. Hoffmann, M., Hilton-Taylor, C., Angulo, A., Böhm, M., Brooks, T.M., Butchart, S.H.M., Carpenter, K.E., Chanson, J., Collen, B., Cox, N.A., 2010. The impact of conservation on the status of the world’s vertebrates. Science 330, 1503–1509. Hurlbert, A.H., Jetz, W., 2007. Species richness, hotspots, and the scale dependence of range maps in ecology and conservation. Proceedings of the National Academy of Sciences 104, 13384–13389. Jablonski, D., Roy, K., Valentine, J.W., 2006. Out of the Tropics: Evolutionary Dynamics of the Latitudinal Diversity Gradient. Science 314, 102–106. Jetz, W., Fine, P.V.A., 2012. Global Gradients in Vertebrate Diversity Predicted by Historical Area-Productivity Dynamics and Contemporary Environment. PLoS Biology 10, e1001292. Jetz, W., Thomas, G.H., Joy, J.B., Hartmann, K., Mooers, A.O., 2012. The global diversity of birds in space and time. Nature 491, 444–448. Jønsson, K.A., Fabre, P.-H., Ricklefs, R.E., Fjeldså, J., 2011. Major global radiation of corvoid birds originated in the proto-Papuan archipelago. Proc Natl Acad Sci USA 108, 2328–2333. Kier, G., Kreft, H., Lee, T.M., Jetz, W., Ibisch, P.L., Nowicki, C., Mutke, J., Barthlott, W., 2009. A global assessment of endemism and species richness across island and mainland regions. Proc Natl Acad Sci USA 106, 9322– 9327. Lijtmaer, D.A., Kerr, K.C.R., Barreira, A.S., Hebert, P.D.N., Tubaro, P.L., 2011. DNA Barcode Libraries Provide Insight into Continental Patterns of Avian Diversification. PLoS ONE 6, e20744.
- 18 Lovette, I.J., Pérez-Emán, J.L., Sullivan, J.P., Banks, R.C., Fiorentino, I., Córdoba-Córdoba, S., Echeverry-Galvis, M., Barker, F.K., Burns, K.J., Klicka, J., Lanyon, S.M., Bermingham, E., 2010. Molecular Phylogenetics and Evolution. Molecular Phylogenetics and Evolution 57, 753–770. Marques, A.C., Maronna, M.M., Collins, A.G., 2013. Putting GenBank data on the map. Science 341, 1341. Mccain, C.M., 2009. Global analysis of bird elevational diversity. Global Ecology and Biogeography 18, 346–360. Morlon, H., Potts, M.D., Plotkin, J.B., 2010. Inferring the Dynamics of Diversification: A Coalescent Approach. PLoS Biol 8, e1000493. Moyle, R.G., Andersen, M.J., Oliveros, C.H., Steinheimer, F.D., Reddy, S., 2012. Phylogeny and Biogeography of the Core Babblers (Aves: Timaliidae). Systematic Biology 61, 631–651. Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G., Kent, J., 2000. Biodiversity hotspots for conservation priorities. Nature 403, 853–858. Oliveros, C.H., Moyle, R.G., 2010. Origin and diversification of Philippine bulbuls. Molecular Phylogenetics and Evolution 54, 822–832. Oliveros, C.H., Reddy, S., Moyle, R.G., 2012. The phylogenetic position of some Philippine “babblers” spans the muscicapoid and sylvioid bird radiations. Molecular Phylogenetics and Evolution 1–6. Olsson, U., Alström, P., Ericson, P.G.P., Sundberg, P., 2005. Non-monophyletic taxa and cryptic species--Evidence from a molecular phylogeny of leafwarblers (Phylloscopus, Aves). Molecular Phylogenetics and Evolution 36, 261–276. Orme, C., Davies, R.G., Burgess, M., Eigenbrod, F., Pickup, N., Olson, V., Webster, A., Ding, T.-S., Rasmussen, P., Ridgely, R., 2005. Global hotspots of species richness are not congruent with endemism or threat. Nature 436, 1016–1019. Orme, C.D.L., Davies, R.G., Olson, V.A., Thomas, G.H., Ding, T.-S., Rasmussen, P.C., Ridgely, R.S., Stattersfield, A.J., Bennett, P.M., Owens, I.P.F., Blackburn, T.M., Gaston, K.J., 2006. Global patterns of geographic range size in birds. PLoS Biol 4, e208. Phillimore, A.B., Price, T.D., 2008. Density-Dependent Cladogenesis in Birds. PLoS Biol 6, e71. Pybus, O.G., Harvey, P.H., 2000. Testing macro-evolutionary models using incomplete molecular phylogenies. Proceedings of the Royal Society B: Biological Sciences 267, 2267–2272. Reddy, S., Cracraft, J., 2007. Old World Shrike-babblers (Pteruthius) belong with New World Vireos (Vireonidae). Molecular Phylogenetics and Evolution 44, 1352–1357. Reddy, S., Dávalos, L.M., 2003. Geographical sampling bias and its implications for conservation priorities in Africa. Journal of Biogeography 30, 1719–1727. Reddy, S., Driskell, A., Rabosky, D.L., Hackett, S.J., Schulenberg, T.S., 2012. Diversification and the adaptive radiation of the vangas of Madagascar. Proceedings of the Royal Society B: Biological Sciences 279, 2062–2071. Ricklefs, R.E., 2006. Global variation in the diversification rate of passerine birds.
- 19 Ecology 87, 2468–2478. Sanderson, M.J., McMahon, M.M., Steel, M., 2010. Phylogenomics with incomplete taxon coverage: the limits to inference. BMC Evolutionary Biology 10, 155. Siler, C.D., Dececchi, T.A., Merkord, C.L., Davis, D.R., Christiani, T.J., Brown, R.M., 2014. Cryptic diversity and population genetic structure in the rare, endemic, forest-obligate, slender geckos of the Philippines. Molecular Phylogenetics and Evolution 70, 204–209. Thomas, G.H., Orme, C.D.L., Davies, R.G., Olson, V.A., Bennett, P.M., Gaston, K.J., Owens, I.P.F., Blackburn, T.M., 2008. Regional variation in the historical components of global avian species richness. Global Ecology and Biogeography 17, 340–351. Tobias, J.A., Bates, J.M., Hackett, S.J., Seddon, N., 2008. Comment on “The latitudinal gradient in recent speciation and extinction rates of birds and mammals.” Science 319, 901c. Voelker, G., Outlaw, R.K., Reddy, S., Tobler, M., Bates, J.M., Hackett, S.J., Kahindo, C., Marks, B.D., Kerbis Peterhans, J.C., Gnoske, T.P., 2010. A new species of boubou (Malaconotidae: Laniarius) from the Albertine Rift. The Auk 127, 678–689. Weir, J.T., Schluter, D., 2007. The latitudinal gradient in recent speciation and extinction rates of birds and mammals. Science 315, 1574–1576. Willig, M.R., Kaufman, D.M., Stevens, R.D., 2003. Latitudinal gradients of biodiversity: pattern, process, scale, and synthesis. Annu. Rev. Ecol. Evol. Syst. 34, 273–309.
Figure Legends
Figure 1. Maps showing a) total species richness, b) number of NoG (no genetic data) species, and c) proportion of NoG species. Grid cells are 5 x 5 degrees.
Figure 2. Distributions of range sizes for NoG (in gray) and WiG (in black) species. The means and variances are significantly different (P<0.0001).
Figure 3. Relationship between a) clade size and completeness (correlation coefficient=-0.229), b) completeness and data decisiveness (correlation
- 20 coefficient=0.402), and c) clade size and data decisiveness (correlation coefficient=0.025).
Figure 1
(a) Total number of species 0 - 31 32 - 106 107 - 190 191 - 297 298 - 444 445 - 613 614 - 859 860 - 1400
(b) Number of species with no genetic data 0 - 50 51 - 100 101 - 150 151 - 200 201 - 250 251 - 310
(c) Proportion of species with no genetic data 0 - 0.1 0.11- 0.2 0.21- 0.3 0.31- 0.4 0.41- 0.5
Figure 2
0.25
0.2
percentage of species
0.15
0.1
0.05
0
-5
-4
-3
-2
-1
0
log10 area
1
2
3
4
Figure 3
1.0
Completeness
0.8 0.6 0.4 0.2 0 0
100
200
300
400
500
0.8
1.0
400
500
Clade Size
A)
1.0
Data Decisiveness
0.8 0.6 0.4 0.2 0 0
0.2
B)
0.4
0.6
Completeness
1.0
Data Decisiveness
0.8 0.6 0.4 0.2 0 0
C)
100
200
300
Clade Size
- 21 Table 1. Low and High outliers of genetic sampling completeness for the crown clades from Jetz et al. (2012). a) Clades with 50% or less completeness of genetic sampling.
Clade Bucconidae Galbulidae Odontophoridae Indicatoridae Cisticolas, Allies Rallidae Weavers, Allies Apodidae Phalacrocoracidae Larks Antpittas Columbidae Pteroclidae Turnicidae Sunbirds, Flowerpeckers Vireos, Allies Leafbirds, Fairy-Bluebirds Paleognathae Sparrows, Snowfinches, Allies Wagtails, Pipits Whiteyes, Babblers I, Parrotbills Honeyeaters Phoeniculidae
Clade Size 37 20 35 19 148 139 118 105 35 93 51 308 18 18 174 60 15 60
Completeness 0.22 0.30 0.31 0.32 0.32 0.34 0.36 0.38 0.40 0.43 0.43 0.44 0.44 0.44 0.47 0.47 0.47 0.47
Range New World New World New World Old World Old World cosmopolitan Old World cosmopolitan cosmopolitan Old World, Australia New World cosmopolitan Old World Old World, Australia Old World New World,Old World Old World cosmopolitan
T T T T T
Passerine/ NonPasserine NP NP NP NP P NP P NP NP P P NP NP NP P P P NP
41 63
0.49 0.49
Old World cosmopolitan
T
P P
368 184 10
0.49 0.50 0.50
Old World Australasian Old World
T T T
P P NP
T T T T T
T T
a) Clades with 90% or more completeness of genetic sampling.
Clade Icterids Parulidae Trogonidae Hirundinidae Musophagidae Otididae Aegothelidae Anserinae Orthonychidae, Pomatostomidae Brachypteraciidae Cathartidae
Clade Size 104 125 45 85 25 27 11 27
Completeness 0.90 0.90 0.91 0.93 0.96 0.96 1.00 1.00
10 7 9
1.00 1.00 1.00
Range (T = tropical) New World New World cosmopolitan T cosmopolitan Old World T Old World Australasia T cosmopolitan Australasia Old World New World
T T
Passerine/ NonPasserine P P NP P NP NP NP NP P NP NP
- 22 -
Whistlers-part Dendrocygninae Fregatidae Gaviidae Gruidae Bombycillidae, Aliies Nyctibiidae Promeropidae Calcariidae-part Phaethontidae Phoenicopteridae Psophidae Regulus Spheniscidae Sulidae Todidae
5 11 7 7 17 12 9 6 8 5 8 5 8 20 12 7
1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
Old World, Australasia T cosmopolitan T cosmopolitan Holarctic cosmopolitan cosmopolitan New World T Old World T cosmopolitan cosmopolitan cosmopolitan New World T Holarctic Southern Hemisphere cosmopolitan New World T
P NP NP NP NP P NP P P NP NP NP NP NP NP NP
Proportion of species with no genetic data 0 - 0.1 0.11- 0.2 0.21- 0.3 0.31- 0.4 0.41- 0.5
- 23 Highlights • There are disparities of genetic sampling across latitudes and hemispheres • Tropical regions have substantially fewer species with genetic data • The Old World Tropics and Australasia are disproportionately understudied • Global diversification rates need to be re-evaluated in light of sampling biases • Understudied regions in critical need of modern systematic analyses are revealed
S1055-7903(14)00151-1 http://dx.doi.org/10.1016/j.ympev.2014.04.023 YMPEV 4888
To appear in:
Molecular Phylogenetics and Evolution
Received Date: Revised Date: Accepted Date:
1 December 2013 18 April 2014 18 April 2014
Please cite this article as: Reddy, S., What’s missing from avian global diversification analyses?, Molecular Phylogenetics and Evolution (2014), doi: http://dx.doi.org/10.1016/j.ympev.2014.04.023
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
-1What’s missing from avian global diversification analyses? Sushma Reddy Biology Department, Loyola University Chicago, Chicago, IL 60660, USA; [email protected]
-2Abstract The accumulation of vast numbers of molecular phylogenetic studies has contributed to huge knowledge gains in the evolutionary history of birds. This permits subsequent analyses of avian diversity, such as how and why diversification varies across the globe and among taxonomic groups. However, available genetic data for these meta-analyses are unevenly distributed across different geographic regions and taxonomic groups. To comprehend the impact of this variation on the interpretation of global diversity patterns, I examined the availability of genetic data for possible biases in geographic and taxonomic sampling of birds. I identified three main disparities of sampling that are geographically associated with latitude (temperate, tropical), hemispheres (East, West), and range size. Tropical regions, which host the vast majority of species, are substantially less studied. Moreover, Eastern regions, such as the Old World Tropics and Australasia, stand out as being disproportionately undersampled, with up to half of communities not being represented in recent studies. In terms of taxonomic discrepancies, a majority of genetically undersampled clades are exclusively found in tropical regions. My analysis identifies several disparities in the key regions of interest of global diversity analyses. Differential sampling can have considerable impacts on these global comparisons and call into question recent interpretations of latitudinal or hemispheric differences of diversification rates. Moreover, this review pinpoints understudied regions whose biota are in critical need of modern systematic analyses.
-3Keywords: Biodiversity; Birds; Diversity Patterns; DNA sequences; Sampling bias; Taxonomy
1.1 Introduction There is a critical need to understand global patterns of biodiversity to increase knowledge of ecology and evolutionary history and to predict the effects of anthropogenic impacts (Gaston, 2000; Hurlbert and Jetz, 2007; Kier et al., 2009; Myers et al., 2000). A major goal of global biodiversity studies is to examine how and why diversity varies across the world. These studies range from describing diversity patterns such as taxonomic richness and endemism to investigating why some regions might produce and/or accumulate species at different rates (Hawkins et al., 2007; Jetz and Fine, 2012; Jetz et al., 2012; Mccain, 2009; Orme et al., 2006). Examining diversification by estimating speciation and extinction rates can help to understand why diversity varies across different landscapes. For instance, a longstanding question is why the tropics are more diverse than temperate regions (Jablonski et al., 2006; Weir and Schluter, 2007; Willig et al., 2003) and much of the debate has been about whether this is due to the tropics acting as cradles (producing many new species) or museums (supporting the accumulation of many species). Examining and testing hypotheses of global patterns relies on having sufficient comparative data. Birds, with about 10,000 described species (Clements et al., 2012; Dickinson et al., 2003), are generally well-studied using modern systematic tools (i.e. with genetic, morphological, and/or distributional
-4data). The accumulation of genetic data, with roughly two-thirds of all described avian species having some published DNA sequence data, has prompted many recent studies to examine large-scale patterns of avian diversification to test hypotheses such as the latitudinal gradient (Hawkins et al., 2007; Jetz et al., 2012; Lijtmaer et al., 2011; Ricklefs, 2006; Weir and Schluter, 2007), East-West hemispheric differences (Jetz et al., 2012; Lijtmaer et al., 2011), densitydependent diversification rates (Morlon et al., 2010; Phillimore and Price, 2008), and ecological correlates of diversity (Orme et al., 2006; Thomas et al., 2008). Furthermore, the vast accumulation of genetic data enabled a recent study to build a 'complete' phylogeny of all living birds (Jetz et al., 2012), filling in missing species by constraining priors using taxonomy. This allowed the study to analyze diversification rates by representing all extant lineages but implicitly assumes that the contribution of evolutionary rates from missing species is minimal. Jetz et al. (2012) used their phylogeny to study patterns of avian diversification across geographic space and evolutionary time. One major conclusion drawn by the authors was that the disparity in diversification rates was not so profound when comparing temperate and tropical regions, but rather that there was a significant difference in rates between the Eastern and Western Hemispheres or Old and New Worlds (Jetz et al., 2012). Furthermore, they allude to an intriguing pattern showing that "avian assemblages in Australia, Southeast Asia, Africa, and Madagascar are characterized by particularly low average rates" of diversification compared to the global mean (Jetz et al., 2012, p. 445). Here, I examine if there is a bias in terms of availability of genetic data from
-5different regions that might mislead such global analyses. Using taxonomy as a proxy for phylogenetic relationships is not an appropriate assumption for many reasons. Taxonomic classifications in birds have been demonstrated to be exceedingly inaccurate in almost every avian molecular systematics study. Avian taxonomic groups at all levels – orders, suborders, families, subfamilies, genera, species, and subspecies – are not stable (e.g., Alström et al., 2006; Barker et al., 2004; Beresford et al., 2005; Cibois and Cracraft, 2004; Hackett et al., 2008; Oliveros et al., 2012; Olsson et al., 2005; Reddy and Cracraft, 2007). Extreme examples of misclassification include species classified based on phenotypic similarities that are now shown to be in highly divergent lineages, in some cases on the order of families (Alström et al., 2014; Cibois, 2003; Cibois et al., 2001; Reddy and Cracraft, 2007). Moreover, an increasing number of studies point out that even shallow evolutionary divergences are complicated with numerous genera and species not being monophyletic (Burns et al., 2014; Lovette et al., 2010; Moyle et al., 2012; Olsson et al., 2005) or recognized (Siler et al., 2014; Tobias et al., 2008). Calculations of diversification rates can be significantly impacted when these lineages, both deeply divergent and shallow branches, go unrecognized. While diversification analyses such as the Jetz et al. (2012) paper are useful and will no doubt lead to subsequent studies, they must be interpreted in light of how well biodiversity is sampled across the globe. This study is not intended to be a critique of these studies or their methods. Rather, I am interested in using available genetic data to examine how much we know about
-6avian species from different regions of the world. Differential knowledge of regional biodiversity, which is impacted by the feasibility/logistics of acquiring samples and scientific interest, could influence the resulting patterns of global analyses. In all of these studies, the implicit assumption is that missing species will not contribute any different information from that derived from available data. Yet this is true only if missing species are randomly distributed, minimal, or contribute the same signal as species with data. Although a few methods have been proposed to accommodate incomplete sampling for estimating diversification rates (Brock et al., 2011; Pybus and Harvey, 2000; Reddy et al., 2012), simulation studies show that results can be unreliable when clade sampling falls below 75-80% completeness (Brock et al., 2011; Cusimano and Renner, 2010) and varied whether or not missing taxa were randomly distributed. The effect of missing species in the avian phylogeny can only be truly tested when they are, in time, sequenced and included in these comparative studies. Examination of possible biases in sampling is key to understanding the limitations of the results and interpretation of diversification studies. Our knowledge of biodiversity patterns can be biased by both geographic and taxonomic sampling (Boakes et al., 2010; Brock et al., 2011; Cusimano and Renner, 2010; Reddy and Dávalos, 2003; Tobias et al., 2008). Here, I use genetic data as an indicator of whether a species has been examined in a recent systematic study to assess if differences in research effort across the globe are substantial enough to impact studies of diversification patterns. I examined the current state of knowledge of bird phylogeny for sampling biases in terms of
-7geography and taxonomy. Birds are generally better studied that other taxonomic groups, therefore it is likely that the patterns uncovered here are similar or more pronounced in less-studied groups.
1.2 Methods I used the genetic data compiled by Jetz et al., (2012) as indicative of available nucleotide sequences for birds (as of April 2011). Although several recent large-scale phylogenies have been published since, (e.g., Barker et al., 2013; Derryberry et al., 2011; Moyle et al., 2012), my assumption here is that the proportional geographic representation is still similar. Jetz et al. recognized 9,993 species, of which 6,663 had some published genetic data. I categorized all bird species into binary distinctions of with (WiG) or without (NoG) genetic data. Unfortunately, there is no geographic data associated with Genbank accessions, a limitation that still needs to be addressed (Marques et al., 2013). Since linking geographic coordinates to each sample was not possible, I made the simple assumption that a species could have been sampled anywhere across its geographic range. For this analysis, I did not distinguish how much data were available for each taxon, but rather considered species as WiG if they had at least one sequence published. While this might bias the analysis in some ways, for instance species with only one sequence are treated the same as other species for which there are multiple sequences, a more precise analysis is unfortunately not feasible at present given the limitations of Genbank deposits. The Jetz et al. study focused on sequences from 15 commonly used loci. While
-8some species might be represented in GenBank with sequences from other loci only, these are a small minority and would not be useful in a comparative context of a larger phylogeny.
1.2.1 Geographic sampling To examine geographic biases in sampling, I used a dataset of species distributions of all birds assembled and distributed through Birdlife International and NatureServe (BirdLife International, Inc. and NatureServe, 2012) to geographically plot the distribution of WiG and NoG species. In ArcGIS (v.10; Environmental Systems Research Institute, Redlands, CA.), I used the native, resident or breeding distributions and combined shapefiles of species with disjunct distributions (using the ‘Dissolve’ function). I then created a 5° by 5° grid across globe (‘Create Fishnet’ function) and counted the total WiG and NoG species in each grid cell (‘Spatial Join’ and ‘Intersect’ functions). Grid cells of different sizes and projections should still yield similar results since the main comparison of the analysis is the proportion of missing species per area. I focused on grid cells that incorporated terrestrial regions by excluding cells with <30 total species, a reasonable assumption because even remote islands usually host substantially more species. I further examined the relationship of range size to genetic sampling to test if NoG species are more likely to have small ranges and therefore were presumably harder to sample. Taking the total area of each species’ resident or breeding range polygon, I compared the frequency distribution of range sizes for
-9WiG and NoG species. I used an unpaired t-test with Welch’s correction and F test to statistically assess if the means and variances of the two distributions were significantly different using Prism (v. 5; GraphPad Software, San Diego, CA).
1.2.2 Taxonomic sampling To examine taxonomic skewness of genetic data, I used the 129 crown clades in the backbone of the Jetz phylogeny to examine if less sampled groups tended to be from certain geographic regions. I also examined whether there was a relationship between clade size, genetic completeness (i.e. percent of WiG taxa), and average data decisiveness (DD; see Jetz et al., 2012), which might skew the results. For each pairwise comparison, I calculated the correlation coefficient in the R statistical package. While some of these potential issues were inspected by Jetz et al., I examined these correlations to further test the impact of missing species. A negative correlation between clade size and completeness would indicate that complete sampling is harder for larger clades. Data decisiveness, a measure of the ability of incomplete data matrices to resolve trees (Sanderson et al., 2010), was measured as the proportion of branches that the data could resolve across 1000 random trees (Jetz et al. 2012). A relationship between DD and completeness indicates that missing data impacts the determination of phylogenetic relationships and an relationship between DD and clade size could indicate that the number of taxa influences how branches are resolved.
- 10 -
1.3 Results Figure 1 shows maps of the geographic distribution of total species richness, number of NoG species, and proportion of NoG species (number of NoG species divided by total species) in each 5x5 degree grid cell. Global patterns of species richness are similar to those discussed previously (Orme et al., 2005), with highest richness being in the tropical montane regions (i.e. Andes, Eastern Africa, Himalayas). Areas with most NoG species coincide with areas of highest richness. However, when examining the proportion of NoG species, the pattern is remarkably different. Areas with proportionally higher missing data—up to 50% NoG species—are concentrated in the Old World tropics and Australasia. Species range sizes of birds span several orders of magnitude. Figure 2 shows the distribution of range sizes for NoG and WiG species. While they both extend over the same breadth, the means and variances of these categories are significantly different. A higher percentage of NoG species have smaller ranges than WiG species. Figure 3 shows the relationship between completeness, clade size, and data decisiveness. The 129 major clades ranged in size from 5 to 460 species. Within these groups, completeness or percentage of WiG species varied from 21.6% to 100%. Data decisiveness, a measure of the impact of missing data on recovering relationships calculated as the proportion of distinguished branches across 1000 trees (Jetz et al., 2012; Sanderson et al., 2010), spanned from 0.23
- 11 to 1 in these clades. There was no correlation between clade size and average decisiveness (correlation coefficient=0.025), a weak negative correlation between clade size and completeness (-0.229), and a moderate correlation between completeness and average decisiveness (0.402). Table 1 shows the clades at the low and high ends of completeness. Table 1a shows clades with 50% or less completeness. These 23 clades (or 17% of all clades) encompass more than 36% of the total number of NoG species. Of these, 9 clades are exclusively distributed in the Old World, 4 are only in the New World, 1 in Australasia, and 9 are widespread. Furthermore, 16 of the 23 clades predominately occur in tropical regions, while the others generally occur in both tropical and temperate zones. These clades were about evenly spread between the distinctions of passerines and non-passerines. Table 1b shows clades with more than 90% completeness. These are 27 clades that are generally smaller (20 clades are less than or equal to 20 species) with a majority being nonpasserines. These clades are almost evenly distributed across all regions.
1.4 Discussion The increasing number of phylogenetic studies has made a near-complete avian tree of life tenable. Yet efforts of modern biodiversity studies, as measured by genetic analyses, show substantial differences across the globe. In this study, I have identified three main disparities associated with genetic sampling of birds across the world, including between high and low latitudes, among Eastern and Western hemispheres, and across small and large ranges. Given the drastic
- 12 variation in sampling, interpretations of meta-analyses have to be taken in light of the limitations of the data used therein. Many of the disparities in sampling effort have crucial implications for the regions of interest in most analyses of global diversity patterns. Not surprisingly, regions with highest species diversity also have the most NoG species. When looking at both the total number of species as well as the proportion of the community that is NoG, there is substantial disparity across latitudes. Temperate regions, especially North America and Europe, comprise the most studied avifaunas despite having lower species richness. The tropics are poorly sampled in genetic studies. These areas are highlighted as having overall high numbers of NoG species (Fig. 1) and, additionally, the majority of clades with less than 50% completeness are found primarily in the tropics (Table 1). Within the tropics, there is a clear discrepancy between sampling in the Western and Eastern Hemispheres. The New World is better studied than other tropical regions. Some parts of the Old World and Australasia are so poorly known that up to half the communities in these regions are missing from recent studies. These are exactly the regions in the Jetz et al. analysis that showed much lower average diversification rates than the global mean, calling into question whether these calculations are biased by uneven sampling. The Old World and Australasia, in contrast to the Neotropics, are logistically more heterogeneous and have fewer researchers focused on these regions, especially in modern studies. The analysis of NoG and WiG species range sizes implies a connection
- 13 between range size and accessibility. The distribution of WiG species is skewed towards wider ranges, while the NoG species tend to have smaller range sizes. However, the extents of both distributions are similar, indicating there are other factors that impact how biotic surveys and studies are conducted. There are WiG species with extremely restricted ranges and likewise NoG species with very wide ranges (mainly seabirds). Estimates of diversification rates of incompletely sampled clades are known to be unreliable (Brock et al., 2011; Cusimano and Renner, 2010; Pybus and Harvey, 2000). Simulation studies have shown that error rates for diversification estimates increase substantially when clade completeness is less than 75-80% (Brock et al., 2011; Cusimano and Renner, 2010). This would include 80 of the 129 major clades in the Jetz et al. analysis. Although Jetz et al. (2012) compensated for missing species by placing NoG taxa onto the phylogeny by using taxonomy as a proxy for evolutionary history, this still resulted in rates estimates being essentially based on WiG taxa (see their supplementary discussion figure 2). The underlying result is that NoG species do not contribute additional information for diversification rates than those derived from the prior constraints in these Bayesian analyses. Furthermore, my analysis shows genetic sampling completeness is correlated to data decisiveness for phylogenetic relationships within clades (Fig. 3). When taxa are missing phylogenetic information, studies often substitute taxonomy as an indicator of evolutionary history. However, many studies (ironically, using genetic data) have shown that traditional groupings and
- 14 relationships implied by taxonomy are grossly erroneous. These include incorrect assessments of species-limits, monophyly, and phylogenetic relationships, which have further obscured our understanding of biodiversity patterns (for examples see Alström et al., 2011; Barker et al., 2004; Beresford et al., 2005; Fuchs et al., 2006; Hackett et al., 2008; Moyle et al., 2012; Oliveros et al., 2012; Reddy et al., 2012). This underscores the need for increased biodiversity research in understudied regions to clarify these inaccuracies. Additionally, continued surveys of little known regions are also key to identifying new, overlooked or cryptic species (see Athreya, 2006; Voelker et al., 2010).
1.5 Conclusions Differential knowledge across regions impacts interpretations of global biogeographic patterns and has critical implications for the informative power of large-scale meta-analyses. This study shows that there are considerable discrepancies in terms of which areas are well-studied using modern systematic analyses. Many of the under-sampled regions are key areas of focus when addressing big question in global biodiversity. The early biogeographic history of many groups of birds is still debated and some of the understudied regions, such as Asia and Australasia, have been proposed to play a key role in the diversification of some groups (Barker et al., 2004; Cracraft, 2001; Jønsson et al., 2011). There is an urgent need to prioritize studies in these poorly known regions to better understand the evolution of modern birds, both in terms of enumerating true diversity and examining their history of diversification. This is especially very
- 15 important in the most severely understudied regions of the Old World Tropics and Australasia, where areas of high diversity coincide with those most at risk by anthropogenic pressures (Hoffmann et al., 2010; Orme et al., 2005). Addressing these major gaps in our knowledge of avian diversity will require intense efforts to target analyses of these regions. Nevertheless, some of these efforts might be hindered by access to fresh tissue samples or political and cultural impediments to biodiversity surveys. While I am not advocating that genetic studies are the only way to conduct modern systematic analyses, these data do contribute a significant portion towards our overall understanding of diversity. If these key geographic areas continue to be holes in our map of life, we will miss out on a truly comprehensive view of biodiversity.
Acknowledgements This work was conducted with financial support from the US National Science Foundation (DEB-0962078) and facilitated by the National Centre for Biological Sciences (India). Access to species distribution data was generously provided by BirdLife International and NatureServe. I am grateful to D. Treering and V. Varma for advice regarding GIS analyses and J. Bates, L. Davalos, P. Makovicky, R. Nandini, H. Skeen, U. Ramakrishnan, V.V. Robin, and two anonymous reviewers for constructive comments to improve this manuscript.
References Alström, P., Ericson, P.G.P., Olsson, U., Sundberg, P., 2006. Phylogeny and classification of the avian superfamily Sylvioidea. Molecular Phylogenetics
- 16 and Evolution 38, 381–397. Alström, P., Fregin, S., Norman, J.A., Ericson, P.G.P., Christidis, L., Olsson, U., 2011. Multilocus analysis of a taxonomically densely sampled dataset reveal extensive non-monophyly in the avian family Locustellidae. Molecular Phylogenetics and Evolution 58, 513–526. Alström, P., Hooper, D.M., Liu, Y., Olsson, U., Mohan, D., Gelang, M., Le Manh, H., Zhao, J., Lei, F., Price, T.D., 2014. Discovery of a relict lineage and monotypic family of passerine birds. Biology Letters 10, 20131067. Athreya, R., 2006. A new species of Liocichla (Aves: Timaliidae) from Eaglenest Wildlife Sanctuary, Arunachal Pradesh, India. Indian Birds 2, 82–94. Barker, F.K., Burns, K.J., Klicka, J., Lanyon, S.M., Lovette, I.J., 2013. Going to Extremes: Contrasting Rates of Diversification in a Recent Radiation of New World Passerine Birds. Systematic Biology 62, 298–320. Barker, F.K., Cibois, A., Schikler, P., Feinstein, J., Cracraft, J., 2004. Phylogeny and diversification of the largest avian radiation. Proc Natl Acad Sci USA 101, 11040–11045. Beresford, P., Barker, F.K., Ryan, P., Crowe, T.M., 2005. African endemics span the tree of songbirds (Passeri): molecular systematics of several evolutionary “enigmas.” Proceedings of the Royal Society, B 272, 849–858. BirdLife International, Inc., NatureServe, 2012. Bird Species Distribution Maps of the World. Version 2.0. Boakes, E.H., McGowan, P.J.K., Fuller, R.A., Chang-qing, D., Clark, N.E., O'Connor, K., Mace, G.M., 2010. Distorted views of biodiversity: spatial and temporal bias in species occurrence data. PLoS Biology 8, e1000385. Brock, C.D., Harmon, L.J., Alfaro, M.E., 2011. Testing for temporal variation in diversification rates when sampling is incomplete and nonrandom. Systematic Biology 60, 410–419. Burns, K.J., Shultz, A.J., Title, P.O., Mason, N.A., Barker, F.K., Klicka, J., Lanyon, S.M., Lovette, I.J., 2014. Molecular Phylogenetics and Evolution. Molecular Phylogenetics and Evolution 75, 41–77. Clements, J.F., Schulenberg, T.S., Iliff, M.J., Sullivan, B.L., Wood, C.L., 2012. The eBird/Clements checklist of birds of the world: Version 6.7. Downloaded from http://www.birds.cornell.edu/clementschecklist/downloadable-clementschecklist. Cibois, A., 2003. Mitochondrial DNA phylogeny of babblers (Timaliidae). The Auk 120, 35–54. Cibois, A., Cracraft, J., 2004. Assessing the passerine “Tapestry”: phylogenetic relationships of the Muscicapoidea inferred from nuclear DNA sequences. Molecular Phylogenetics and Evolution 32, 264–273. Cibois, A., Slikas, B., Schulenberg, T., Pasquet, E., 2001. An endemic radiation of Malagasy songbirds is revealed by mitochondrial DNA sequence data. Evolution 55, 1198–1206. Cracraft, J., 2001. Avian evolution, Gondwana biogeography and the Cretaceous-Tertiary mass extinction event. Proceedings of the Royal Society, B 268, 459–469. Cusimano, N., Renner, S.S., 2010. Slowdowns in Diversification Rates from Real
- 17 Phylogenies May Not be Real. Systematic Biology 59, 458–464. Derryberry, E.P., Claramunt, S., Derryberry, G., Chesser, R.T., Cracraft, J., Aleixo, A., Pérez-Emán, J., Remsen, J.V., Jr, Brumfield, R.T., 2011. Lineage Diversification and Morphological Evolution in a Large-Scale Continental Radiation: the Neotropical Ovenbirds and Woodcreepers (Aves: Furnariidae). Evolution 65, 2973–2986. Dickinson, E.C., Bahr, N., Dowsett, R., Pearson, D., Remsen, V., Roselaar, C.S., Schodde, D., 2003. The Howard and Moore Complete Checklist of Birds of the World, revised and enlarged. 3rd edition. Christopher Helm, London, UK. Fuchs, J., Fjeldså, J., Bowie, R.C.K., Voelker, G., Pasquet, E., 2006. The African warbler genus Hyliota as a lost lineage in the Oscine songbird tree: molecular support for an African origin of the Passerida. Molecular Phylogenetics and Evolution 39, 186–197. Gaston, K., 2000. Global patterns in biodiversity. Nature 405, 220–227. Hackett, S.J., Kimball, R.T., Reddy, S., Bowie, R.C.K., Braun, E.L., Braun, M.J., Chojnowski, J.L., Cox, W.A., Han, K.-L., Harshman, J., Huddleston, C.J., Marks, B.D., Miglia, K.J., Moore, W.S., Sheldon, F.H., Steadman, D.W., Witt, C.C., Yuri, T., 2008. A phylogenomic study of birds reveals their evolutionary history. Science 320, 1763–1768. Hawkins, B.A., Diniz-Filho, J.A.F., Jaramillo, C.A., Soeller, S.A., 2007. Climate, Niche Conservatism, and the Global Bird Diversity Gradient. American Naturalist 170, S16–S27. Hoffmann, M., Hilton-Taylor, C., Angulo, A., Böhm, M., Brooks, T.M., Butchart, S.H.M., Carpenter, K.E., Chanson, J., Collen, B., Cox, N.A., 2010. The impact of conservation on the status of the world’s vertebrates. Science 330, 1503–1509. Hurlbert, A.H., Jetz, W., 2007. Species richness, hotspots, and the scale dependence of range maps in ecology and conservation. Proceedings of the National Academy of Sciences 104, 13384–13389. Jablonski, D., Roy, K., Valentine, J.W., 2006. Out of the Tropics: Evolutionary Dynamics of the Latitudinal Diversity Gradient. Science 314, 102–106. Jetz, W., Fine, P.V.A., 2012. Global Gradients in Vertebrate Diversity Predicted by Historical Area-Productivity Dynamics and Contemporary Environment. PLoS Biology 10, e1001292. Jetz, W., Thomas, G.H., Joy, J.B., Hartmann, K., Mooers, A.O., 2012. The global diversity of birds in space and time. Nature 491, 444–448. Jønsson, K.A., Fabre, P.-H., Ricklefs, R.E., Fjeldså, J., 2011. Major global radiation of corvoid birds originated in the proto-Papuan archipelago. Proc Natl Acad Sci USA 108, 2328–2333. Kier, G., Kreft, H., Lee, T.M., Jetz, W., Ibisch, P.L., Nowicki, C., Mutke, J., Barthlott, W., 2009. A global assessment of endemism and species richness across island and mainland regions. Proc Natl Acad Sci USA 106, 9322– 9327. Lijtmaer, D.A., Kerr, K.C.R., Barreira, A.S., Hebert, P.D.N., Tubaro, P.L., 2011. DNA Barcode Libraries Provide Insight into Continental Patterns of Avian Diversification. PLoS ONE 6, e20744.
- 18 Lovette, I.J., Pérez-Emán, J.L., Sullivan, J.P., Banks, R.C., Fiorentino, I., Córdoba-Córdoba, S., Echeverry-Galvis, M., Barker, F.K., Burns, K.J., Klicka, J., Lanyon, S.M., Bermingham, E., 2010. Molecular Phylogenetics and Evolution. Molecular Phylogenetics and Evolution 57, 753–770. Marques, A.C., Maronna, M.M., Collins, A.G., 2013. Putting GenBank data on the map. Science 341, 1341. Mccain, C.M., 2009. Global analysis of bird elevational diversity. Global Ecology and Biogeography 18, 346–360. Morlon, H., Potts, M.D., Plotkin, J.B., 2010. Inferring the Dynamics of Diversification: A Coalescent Approach. PLoS Biol 8, e1000493. Moyle, R.G., Andersen, M.J., Oliveros, C.H., Steinheimer, F.D., Reddy, S., 2012. Phylogeny and Biogeography of the Core Babblers (Aves: Timaliidae). Systematic Biology 61, 631–651. Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G., Kent, J., 2000. Biodiversity hotspots for conservation priorities. Nature 403, 853–858. Oliveros, C.H., Moyle, R.G., 2010. Origin and diversification of Philippine bulbuls. Molecular Phylogenetics and Evolution 54, 822–832. Oliveros, C.H., Reddy, S., Moyle, R.G., 2012. The phylogenetic position of some Philippine “babblers” spans the muscicapoid and sylvioid bird radiations. Molecular Phylogenetics and Evolution 1–6. Olsson, U., Alström, P., Ericson, P.G.P., Sundberg, P., 2005. Non-monophyletic taxa and cryptic species--Evidence from a molecular phylogeny of leafwarblers (Phylloscopus, Aves). Molecular Phylogenetics and Evolution 36, 261–276. Orme, C., Davies, R.G., Burgess, M., Eigenbrod, F., Pickup, N., Olson, V., Webster, A., Ding, T.-S., Rasmussen, P., Ridgely, R., 2005. Global hotspots of species richness are not congruent with endemism or threat. Nature 436, 1016–1019. Orme, C.D.L., Davies, R.G., Olson, V.A., Thomas, G.H., Ding, T.-S., Rasmussen, P.C., Ridgely, R.S., Stattersfield, A.J., Bennett, P.M., Owens, I.P.F., Blackburn, T.M., Gaston, K.J., 2006. Global patterns of geographic range size in birds. PLoS Biol 4, e208. Phillimore, A.B., Price, T.D., 2008. Density-Dependent Cladogenesis in Birds. PLoS Biol 6, e71. Pybus, O.G., Harvey, P.H., 2000. Testing macro-evolutionary models using incomplete molecular phylogenies. Proceedings of the Royal Society B: Biological Sciences 267, 2267–2272. Reddy, S., Cracraft, J., 2007. Old World Shrike-babblers (Pteruthius) belong with New World Vireos (Vireonidae). Molecular Phylogenetics and Evolution 44, 1352–1357. Reddy, S., Dávalos, L.M., 2003. Geographical sampling bias and its implications for conservation priorities in Africa. Journal of Biogeography 30, 1719–1727. Reddy, S., Driskell, A., Rabosky, D.L., Hackett, S.J., Schulenberg, T.S., 2012. Diversification and the adaptive radiation of the vangas of Madagascar. Proceedings of the Royal Society B: Biological Sciences 279, 2062–2071. Ricklefs, R.E., 2006. Global variation in the diversification rate of passerine birds.
- 19 Ecology 87, 2468–2478. Sanderson, M.J., McMahon, M.M., Steel, M., 2010. Phylogenomics with incomplete taxon coverage: the limits to inference. BMC Evolutionary Biology 10, 155. Siler, C.D., Dececchi, T.A., Merkord, C.L., Davis, D.R., Christiani, T.J., Brown, R.M., 2014. Cryptic diversity and population genetic structure in the rare, endemic, forest-obligate, slender geckos of the Philippines. Molecular Phylogenetics and Evolution 70, 204–209. Thomas, G.H., Orme, C.D.L., Davies, R.G., Olson, V.A., Bennett, P.M., Gaston, K.J., Owens, I.P.F., Blackburn, T.M., 2008. Regional variation in the historical components of global avian species richness. Global Ecology and Biogeography 17, 340–351. Tobias, J.A., Bates, J.M., Hackett, S.J., Seddon, N., 2008. Comment on “The latitudinal gradient in recent speciation and extinction rates of birds and mammals.” Science 319, 901c. Voelker, G., Outlaw, R.K., Reddy, S., Tobler, M., Bates, J.M., Hackett, S.J., Kahindo, C., Marks, B.D., Kerbis Peterhans, J.C., Gnoske, T.P., 2010. A new species of boubou (Malaconotidae: Laniarius) from the Albertine Rift. The Auk 127, 678–689. Weir, J.T., Schluter, D., 2007. The latitudinal gradient in recent speciation and extinction rates of birds and mammals. Science 315, 1574–1576. Willig, M.R., Kaufman, D.M., Stevens, R.D., 2003. Latitudinal gradients of biodiversity: pattern, process, scale, and synthesis. Annu. Rev. Ecol. Evol. Syst. 34, 273–309.
Figure Legends
Figure 1. Maps showing a) total species richness, b) number of NoG (no genetic data) species, and c) proportion of NoG species. Grid cells are 5 x 5 degrees.
Figure 2. Distributions of range sizes for NoG (in gray) and WiG (in black) species. The means and variances are significantly different (P<0.0001).
Figure 3. Relationship between a) clade size and completeness (correlation coefficient=-0.229), b) completeness and data decisiveness (correlation
- 20 coefficient=0.402), and c) clade size and data decisiveness (correlation coefficient=0.025).
Figure 1
(a) Total number of species 0 - 31 32 - 106 107 - 190 191 - 297 298 - 444 445 - 613 614 - 859 860 - 1400
(b) Number of species with no genetic data 0 - 50 51 - 100 101 - 150 151 - 200 201 - 250 251 - 310
(c) Proportion of species with no genetic data 0 - 0.1 0.11- 0.2 0.21- 0.3 0.31- 0.4 0.41- 0.5
Figure 2
0.25
0.2
percentage of species
0.15
0.1
0.05
0
-5
-4
-3
-2
-1
0
log10 area
1
2
3
4
Figure 3
1.0
Completeness
0.8 0.6 0.4 0.2 0 0
100
200
300
400
500
0.8
1.0
400
500
Clade Size
A)
1.0
Data Decisiveness
0.8 0.6 0.4 0.2 0 0
0.2
B)
0.4
0.6
Completeness
1.0
Data Decisiveness
0.8 0.6 0.4 0.2 0 0
C)
100
200
300
Clade Size
- 21 Table 1. Low and High outliers of genetic sampling completeness for the crown clades from Jetz et al. (2012). a) Clades with 50% or less completeness of genetic sampling.
Clade Bucconidae Galbulidae Odontophoridae Indicatoridae Cisticolas, Allies Rallidae Weavers, Allies Apodidae Phalacrocoracidae Larks Antpittas Columbidae Pteroclidae Turnicidae Sunbirds, Flowerpeckers Vireos, Allies Leafbirds, Fairy-Bluebirds Paleognathae Sparrows, Snowfinches, Allies Wagtails, Pipits Whiteyes, Babblers I, Parrotbills Honeyeaters Phoeniculidae
Clade Size 37 20 35 19 148 139 118 105 35 93 51 308 18 18 174 60 15 60
Completeness 0.22 0.30 0.31 0.32 0.32 0.34 0.36 0.38 0.40 0.43 0.43 0.44 0.44 0.44 0.47 0.47 0.47 0.47
Range New World New World New World Old World Old World cosmopolitan Old World cosmopolitan cosmopolitan Old World, Australia New World cosmopolitan Old World Old World, Australia Old World New World,Old World Old World cosmopolitan
T T T T T
Passerine/ NonPasserine NP NP NP NP P NP P NP NP P P NP NP NP P P P NP
41 63
0.49 0.49
Old World cosmopolitan
T
P P
368 184 10
0.49 0.50 0.50
Old World Australasian Old World
T T T
P P NP
T T T T T
T T
a) Clades with 90% or more completeness of genetic sampling.
Clade Icterids Parulidae Trogonidae Hirundinidae Musophagidae Otididae Aegothelidae Anserinae Orthonychidae, Pomatostomidae Brachypteraciidae Cathartidae
Clade Size 104 125 45 85 25 27 11 27
Completeness 0.90 0.90 0.91 0.93 0.96 0.96 1.00 1.00
10 7 9
1.00 1.00 1.00
Range (T = tropical) New World New World cosmopolitan T cosmopolitan Old World T Old World Australasia T cosmopolitan Australasia Old World New World
T T
Passerine/ NonPasserine P P NP P NP NP NP NP P NP NP
- 22 -
Whistlers-part Dendrocygninae Fregatidae Gaviidae Gruidae Bombycillidae, Aliies Nyctibiidae Promeropidae Calcariidae-part Phaethontidae Phoenicopteridae Psophidae Regulus Spheniscidae Sulidae Todidae
5 11 7 7 17 12 9 6 8 5 8 5 8 20 12 7
1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
Old World, Australasia T cosmopolitan T cosmopolitan Holarctic cosmopolitan cosmopolitan New World T Old World T cosmopolitan cosmopolitan cosmopolitan New World T Holarctic Southern Hemisphere cosmopolitan New World T
P NP NP NP NP P NP P P NP NP NP NP NP NP NP
Proportion of species with no genetic data 0 - 0.1 0.11- 0.2 0.21- 0.3 0.31- 0.4 0.41- 0.5
- 23 Highlights • There are disparities of genetic sampling across latitudes and hemispheres • Tropical regions have substantially fewer species with genetic data • The Old World Tropics and Australasia are disproportionately understudied • Global diversification rates need to be re-evaluated in light of sampling biases • Understudied regions in critical need of modern systematic analyses are revealed