PhenoSpace: A Shiny application to visualize trait data in the phenotypic space of the global spectrum of plant form and function

Abstract A recent analysis of variation in six major traits conducted on a large worldwide sample of vascular plant species showed that three‐quarters of trait variation was captured by a two‐dimensional global spectrum of plant form and function (“global spectrum” hereafter). We developed the PhenoSpace application, whose aim is to visualize and export the position of any individual/population/species in the phenotypic space of the global spectrum. PhenoSpace is a Shiny application that helps users to manipulate and visualize data pertaining to the global spectrum of plant form and function. It is freely accessible at the following URL: https://shiny.cefe.cnrs.fr/PhenoSpace/. PhenoSpace has three main functionalities. First, it allows users to visualize the phenotypic space of the global spectrum using different combinations of traits and growth forms. Second, trait data from any new user‐defined dataset can be projected onto the phenotypic space of the global spectrum, provided that at least two of the six traits are available. Finally, figures produced and loadings of the imported data on the PCA axes can be downloaded, allowing users to conduct further analyses. PhenoSpace fulfills the practical goal of positioning plants in the phenotypic space of the global spectrum, making it possible to compare trait variation at any level of organization against the worldwide background. This serves a major aim of comparative plant ecology, which is to put specific sets of individuals, populations or species into a broader context, facilitating comparison and synthesis of results across different continents and environments using relevant indicators of plant design and function.


| INTRODUC TI ON
Traits, which are measurable properties of individuals related to their functioning and modulating their fitness (Calow, 1987;McGill et al., 2006;Violle et al., 2007), are key to study biodiversity from a functional perspective (Enquist et al., 2015;Garnier et al., 2016;Lavorel et al., 2007 for reviews). In plants, an established body of research has been devoted to the identification of key traits allowing a synthetic description of phenotypes in a way that is relevant to their functioning and ecology (Grime, 1977;Laughlin, 2014;Weiher et al., 1999;Westoby, 1998). A recent study conducted on the largest dataset ever compiled of six major traits critical to growth, survival, and reproduction and measured on a sample of vascular plant species distributed worldwide, showed that three-quarters of trait variation could be captured in a two-dimensional "global spectrum of plant form and function" ("global spectrum" hereafter: Díaz et al., 2016). One major dimension reflects the size of whole plants and their parts, while the other represents the leaf economics spectrum (Wright et al., 2004), which runs from quick to slow return on investments of nutrients or dry mass in leaves. As a major advance in our understanding of plant phenotypes, this global spectrum provides a backdrop for describing species from a functional perspective using relevant axes of variation (discussed in Díaz et al., 2016;Westoby et al., 2002).
We, therefore, expect extensive use of this phenotypic space as a reference in plant trait-based research (907 citations to the Díaz et al., 2016 paper in Google Scholar as of 4 August 2020). In this context, we developed the application PhenoSpace, which fulfills the practical goal of positioning plants in the phenotypic space of the global spectrum, making it possible to compare trait variation at any level of organization against the worldwide background.
This serves a major aim of comparative plant ecology, which is to put specific sets of individuals, populations or species into a broader context, facilitating comparison and synthesis of results across different continents and environments using relevant functional axes. More precisely, PhenoSpace allows one to: (a) visualize the outcome of various multivariate analyses run with the original dataset compiled by Díaz et al. (2016) or subsets thereof; (b) project and visualize any user-defined dataset in the phenotypic space of the global spectrum; and (c) download figures produced by the user and/or coordinates of the imported dataset on the PC axes of plant size and the leaf economics spectrum.

| THE FUN C TI ONALITIE S OF PH EN OS PACE
Shiny is an R package that makes it easy to build interactive web applications from R (Chang et al., 2019). PhenoSpace, the Shiny application presented here, is designed to help users manipulate and visualize data pertaining to the global spectrum. It is available at the following URL: https://shiny.cefe.cnrs.fr/Pheno Space/. Its functionalities, accessible from three different tabs, are described below.
TA B L E 1 Traits involved in the phenotypic space of the global spectrum of plant form and function (Díaz et al., 2016) and their functional significance Pérez-Harguindeguy et al., 2013

| Visualizing the phenotypic space of the global spectrum
The first functionality of PhenoSpace ("Customize the PCA" tab) is to display the outcomes of multivariate analyses based on the original dataset explored by Díaz et al. (2016).
Six traits are involved in the definition of the phenotypic space of the global spectrum (Table 1 for

| Plotting data onto the phenotypic space of the global spectrum
The second functionality of PhenoSpace ("Project your data" tab) allows the projection of new data from an uploaded file onto the user-defined phenotypic space of the global spectrum. Therefore, any dataset providing values for at least two of the six traits can be easily compared with the worldwide trait distribution.
The first step is to prepare a csv file that contains the data to be projected. Trait values should be organized in columns whose headings correspond to the abbreviations of trait names spelt out exactly as shown in Table 1. Trait values should be expressed in the units shown in Table 1, and no data transformation is required before the upload. Each line of the dataset can describe an individual, a population, a species or even a community ("entity" hereafter). Entities with missing values (coded NA) are excluded from the analysis. No data completion or gap-filling method is performed in the application. If it is considered useful for a specific dataset, it should be performed by the user before uploading the file. Extracolumns are allowed and can subsequently be used to identify subsets of data on the plot (see below).
The following conditions must be verified on the application set-

| Download figures and tables
The third tab (Downloads) allows the user to save the figures in a png, pdf, or svg format specifying the image size and resolution (for We found that early-successional species had lower values than latesuccessional species on the size axis (PC1) while no difference between stages was found on the leaf economics axis (PC2).

| D ISCUSS I ON AND CON CLUS I ON
The This can be done not only for species, but also for populations or genotypes within species, or even for higher taxonomic ranks or communities. PhenoSpace is a service to the community, which is in principle similar to, for example, the taxonomic name resolution service (TNRS, Boyle et al., 2013)  In these cases, one can consider that the use of large datasets such as the original dataset explored by Díaz et al. (2016) is more appropriate to construct a meaningful multivariate space. PhenoSpace relies on this idea as imported data are considered as inactive entities in the PCA. Finally, there is an increasing interest in the description of volume and shapes of phenotypic hypervolumes (Blonder, 2018). The global spectrum provides a standardized multivariate space for each set of traits, from which hypervolumes of projected data can be computed (using the download functionality, see section 2.3) and easily compared.
PhenoSpace is an open source project (available on a GitHub repository at: https://github.com/jsegr estin/ pheno space). The application will be regularly updated to make sure that it provides the most detailed phenotypic space of plants as our understanding and description of plant phenotype improves. We will therefore work in close collaboration with plant trait databases (e.g., the TRY database, see Kattge et al., 2011Kattge et al., , 2020 to include new species and/or traits in future versions of the application. Our decision rule will be to include a new trait when data are available for a large amount of species in common with the current global spectrum and displays a wide range of variation. The version number, a report detailing the new features, and links to old versions of the application will be available on the webpage. PhenoSpace could also be further developed to position entities in phenotypic spaces already defined with sets of traits different from those of the global spectrum. Two schemes would be particularly interesting in this perspective. The first one is that of F I G U R E 3 Post hoc analysis using the coordinates of the 55 species from Navas et al. (2010) on the global spectrum axes (projection shown in Figure 2). Boxplots represent the 0.25, 0.5, and 0.75 quantiles (colored boxes) and the range of values (whiskers) covered by species in each successional stage. Boxplots that share the same letter are not significantly different: post hoc Tukey test, p value > .05 the leaf economics spectrum (Wright et al., 2004), which is extensively used to position species along an axis of resource acquisition and conservation, and involves the six following leaf traits: photosynthetic and respiration rates, LMA, N mass , leaf phosphorus content per unit mass and leaf life span. The second one is the CSR ecological strategy scheme (Grime, 1977(Grime, , 2001, in which entities are assigned scores on axes characterizing their degree of competitive ability (C), stress-tolerance (S), and ruderality ( The current and future versions of PhenoSpace will serve a major aim of comparative plant ecology, which is to put specific sets of individuals/populations/species/communities into a broader context (Díaz et al., 2016), allowing one to synthesize results across different continents and environments using relevant indicators of plant functioning (cf. Westoby et al., 2002). Future developments may not only be based on existing schemes of phenotypic spaces but could also incorporate other relevant functional dimensions as our understanding of plant phenotypes improves.

ACK N OWLED G M ENTS
We thank the Platform Systèmes d'Information en Ecologie (SIE) of the Centre d'Ecologie Fonctionnelle et Evolutive (CEFE CNRS), Montpellier, France for hosting the application and for the maintenance of the shiny server. JS and KS were supported by fellowships from the Montpellier University. SD was supported by CONICET, FONCyT, IAI, and SECyT-UNC.

CO N FLI C T O F I NTE R E S T
The authors state that there is no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The code of the application is available on a GitHub repository at: https://github.com/jsegr estin/ pheno space. All datasets used in this paper are from published studies. To avoid conflicts of interest, the dataset of the global spectrum will be published soon as a datapaper and a link will be available on the webpage once it is published.