ORIGINAL ARTICLE Root traits of European Vicia faba cultivars—Using machine learning to explore adaptations to agroclimatic conditions Jiangsan Zhao 1 | Peter Sykacek 2 | Gernot Bodner 3 | Boris Rewald 1 1 Department of Forest and Soil Sciences, University of Natural Resources and Life Sciences, Vienna (BOKU), 1190 Tulln an der Donau, Austria 2 Department of Biotechnology, University of Natural Resources and Life Sciences, Vienna (BOKU), 1190 Tulln an der Donau, Austria 3 Division of Agronomy, Department of Crop Sciences, University of Natural Resources and Life Sciences, Vienna (BOKU), 3430 Tulln an der Donau, Austria Correspondence Boris Rewald, Department of Forest and Soil Sciences, University of Natural Resources and Life Sciences, Vienna (BOKU), Peter‐Jordan‐ Straße 82, 1190 Tulln an der Donau, Austria. Email: brewald@rootecology.de Funding information European Union's Seventh Framework Pro- gram, Grant/Award Number: 613781 (EUROLEGUME) Abstract Faba bean (Vicia faba L.) is an important source of protein, but breeding for increased yield stability and stress tolerance is hampered by the scarcity of phenotyping information. Because comparisons of cultivars adapted to different agroclimatic zones improve our understanding of stress tolerance mechanisms, the root architecture and morphology of 16 European faba bean cultivars were studied at maturity. Different machine learning (ML) approaches were tested in their usefulness to analyse trait variations between cultivars. A supervised, that is, hypothesis‐ driven, ML approach revealed that cultivars from Portugal feature greater and coarser but less frequent lateral roots at the top of the taproot, potentially enhancing water uptake from deeper soil horizons. Unsupervised clustering revealed that trait differences between northern and southern cultivars are not predominant but that two cultivar groups, independently from major and minor types, differ largely in overall root system size. Methodological guidelines on how to use powerful ML methods such as random forest models for enhancing the phenotypical exploration of plants are given. KEYWORDS breeding, faba bean (Vicia faba L.), group classification, kernel spectral clustering, k‐nearest neighbour, phenotyping, random forest, root traits selection, supervised learning, unsupervised learning 1 | INTRODUCTION Nearly a century has been spent systematically collecting and preserv- ing the genetic diversity in plants (Westengen, Jeppson, & Guarino, 2013). Seed banks have been established as a source of genetic varia- tion for improving agricultural crops. However, the unprecedented growth of genomic information made the phenotyping bottleneck more apparent than ever: Genomic information cannot be used for breeding without relating it to phenotype information (Fiorani & Schurr, 2013). Root phenotyping is as important as shoot phenotyping, because plant performance strongly depends on its root architecture and function (de Dorlodot et al., 2007; Kirkegaard, Lilley, Howe, & Graham, 2007). Examples are the benefit of shallow rooting in P‐poor soils, maximizing P uptake from the topsoil (Miguel, Widrig, Vieira, Brown, & Lynch, 2013), the strong negative correlation between root respiration and above ground growth under NO 3 /NH 4 ‐nutrition (Rewald, Kunze, & Godbold, 2016), or the advantages of fast root growth and/or deep rooting for extended water uptake (Kirkegaard et al., 2007; Wasson et al., 2012). Although an optimal root system architecture has the potential to improve the resource use efficiency of crops (Lynch, 2007), the variation of root traits is an underexploited resource in plant breeding. Transient drought is expected to increase in frequency and sever- ity as climate change intensifies (Dai, 2013). Cultivars with a high water stress tolerance or the ability to avoid or delay the onset of stress by mechanisms such as earliness or extended water uptake capacities can help to improve yield stability under these future environmental conditions. The screening of genetic resources by phenotyping for plant traits attenuating the consequences of climate change is thus key (Mackay et al., 2004). Phenotypes of cultivated crop accessions -------------------------------------------------------------------------------------------------------------------------------- This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2017 John Wiley & Sons Ltd Received: 29 June 2017 Revised: 18 August 2017 Accepted: 22 August 2017 DOI: 10.1111/pce.13062 Plant Cell Environ. 2017;1–13. wileyonlinelibrary.com/journal/pce 1