Genomic prediction in rye A THESIS SUBMITTED TO THE FACULTY OF AGRICULTURAL SCIENCES OF UNIVERSITY OF HOHENHEIM IN FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DR.SC.AGR. / PH.D. IN AGRICULTURAL SCIENCES Angela Maria Bernal Vasquez FacultyofAgriculturalSciences UniversityofHohenheim March2017 FACULTYOFAGRICULTURALSCIENCE InstituteofCropScience UniversityofHohenheim BiostatisticsUnit Prof. Dr. Hans-PeterPiepho Genomic prediction in rye Dissertation submittedinfulfillmentoftheregulationstoacquirethedegree “DoktorderAgrarwissenschaften” (Dr.sc.agr. /Ph.D.inAgriculturalSciences) tothe FacultyofAgriculturalSciences presentedby: AngelaMariaBernalVasquez bornin: Bogota,Colombia submittedin: March2017 This thesis was accepted as a doctoral thesis (Dissertation) in fulfilment of the regulations toacquirethedoctoraldegree“DoktorderAgrarwissenschaften”bytheFacultyofAgricultural SciencesatUniversityofHohenheimonAugust4th,2017 Dateoftheoralexamination: October24th,2017 ExaminationCommittee Chairpersonoftheoralexamination Prof. Dr. Bennewitz SupervisorandReviewer Prof. Dr. Hans-PeterPiepho Co-Reviewer Prof. Dr. Chris-CarolinScho¨n (TechnischeUniversita¨tMu¨nchen) Additionalexaminer Prof. Dr. AlbrechtE.Melchinger iii To my closest ancestors, my full-sib and my tester i ii Acknowledgments During this long journey of making my PhD I have met people who have walked with me and havehelpedmeeitherwithwords,withfactsorwithenergy. ThefirstoneIamindebtedto,and to whom I will never have enough words to thank is my “Doktorvater” Prof. Dr. Hans-Peter Piepho. He has guided my research with accurate advice and scientific curiosity. Thank you for the endless support and patience, for having always time for answers and for your constant inspiration. Ithasbeenagreathonortoworkwithyou. Many thanks also to Prof. Dr. Chris-Carolin Scho¨n for her always fine-tuned insights as project partner, co-author and now as part of the examination committee. Thanks to Prof. Dr. A. E. Melchinger for his plant breeding lectures building a baseline for this work and also for agreeingtoserveonmydissertationcommittee. Massive dues to Dr. Andres Gordillo, who made the bridge for me towards Germany and had always accurate personal and professional advices. He has contributed with very practical viewsofthesometimesencryptedresultsandhasalwaysallocatedtimetofollow-upanddiscuss ourwork. It was a pleasure for me to be part of the elite group of Rye-Select. I learnt so many things during the project meetings and profited from the discussions and ideas exchanged. Thanks to all the group led by Dr. Peer Wilde and Dr. Viktor Korzun, for allowing the use of the data and the crew support. Big big thanks to my co-authors Malthe Schmidt and Manfred Scho¨nleben (wherever you are) for such good job cooperation and empathy. To Dr. H. F. Utz, who shared allhisPlabStatoutliersecretsandwatchedoutthispublication. Doublethankstomycolleague JensMo¨ringforusefulhintsatthebeginningoftheway,forco-authoringandhisunconditional supportovertheseyears. IamverygratefultomycolleaguesintheBiostatisticsgroup,theoldandthenewones,for iii the cheerful and positive atmosphere, useful discussions and good laughs. To Zeynep Akyildiz for taking care of my logistics and her sweet willingness to help. To Joseph O. Ogutu for his genomic selection input and Waqas Malik for his R guidance, to Vanda Lourenc¸o for her cheerfuladvice,toTorbenSchulz-StreeckmySASsensei,andtoSteffenHadashforwalk&talk company while philosophical and statistical musing. Extra thanks to my Hohenheim friends, the ones that are still there and the ones that have already left, because celebrating success felt double with your company and sharing sadness felt half: Marcelo (my soul buddy), Oscar, Joaqu´ın, Hanna & Juan & Isabel, Manuel, Yudith, Valheria & Jose´, Lina, Lucia & Marco. The thanks also cross the ocean and go to my friends in Colombia (Pipe, Migue, Natali, AngelaP) and to my former professors Jose´ Miguel Cotes and Luis Ernesto Rodr´ıguez. I could always countonthemtosolvecontextualizationdoubtsandtokeepconnectedtomyrootsandspirit. IwanttothankmyGermanfamily,whotookmeintheirheartwithcareandwarmth. Tomy Colombian family, my mother, my father and my brother to whom I dedicate this work. You havebeenmyinfinitesunateverymomentfillingmewithlove,passionand“empuje”: Gracias por su infinito apoyo, a´nimo y energ´ıa!. Thank you to my dearest Gerald, for biking, walking and swimming with me, no matter what the conditions were, hand in hand all throughout this GranFondo. Thankyousoooomuchforever!. Ifinallythankyouunknownreaderforsearchinganswersinmywork. iv Table of Contents Acknowledgments iii Nomenclature ix ListofFigures xix ListofTables xxv 1 Introduction 1 1.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Thehybridryebreedingprogram . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.3 Residualanalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.4 Phenotypicanalysistowardsgenomicprediction . . . . . . . . . . . . . . . . . 5 1.4.1 Astage-wiseapproachforphenotypicandgenomicpredictionanalyses 5 1.4.2 Genomicpredictionandrelatedfactors . . . . . . . . . . . . . . . . . 6 1.4.3 Genomicpredictionmodels . . . . . . . . . . . . . . . . . . . . . . . 7 1.5 Themodelchoice: linkingphenotypestogenotypes . . . . . . . . . . . . . . . 9 1.5.1 Spatialmodelsasanadd-onofthephenotypicmodel . . . . . . . . . . 9 1.5.2 Genomic prediction - cross validation (GP-CV) as a tool for model selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 1.5.3 Accountingforgenotype-by-environment(GE)interactioneffects . . . 11 1.6 Objectivesandhypotheses . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 1.7 Outlineofthethesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 v