A framework for simulations in quantitative genetics and breeding with simplePHENOTYPES
Welcome
Build simulations whose assumptions, genetic mechanisms, and realized outcomes you can explain and check.
simplePHENOTYPES is built for researchers who want to start from their own genotype data, state the genetic assumptions of a simulation explicitly, and check what each assumption does to the outcome. The same framework carries a simulated trait from a single panel of genotypes through crossing, prediction, and selection, so a question can grow from one phenotype to a whole breeding scheme without changing tools.
This book is for researchers in plant, animal, and human genetics with basic R knowledge. Examples follow the tidyverse style, introduce unfamiliar functions, and compare design parameters with simulated results in simple tables. It connects the simplePHENOTYPES interface to the statistical models and computational methods behind it. Worked examples come first; interpretation, assumptions, and theory follow. Longer derivations belong in the appendices.
Every chapter in the chapter plan now has a worked draft, and every exported function and registered method has at least one executed example. The pilot chapters passed independent scientific review. The methods chapters passed a same-model theory review, with every error it found corrected, and still await the independent cross-model review. Example execution and scientific review are separate checks; the evidence register records the status of each claim.
Start with a complete simulation
A first simulation starts with importing your own genotype data and uses bundled observed maize genotypes for a reproducible walkthrough of model specification, diagnostics, and file export. It also explains why the requested variance allocation and the realized variance ratio can differ.
The book serves plant, animal, and human genetics with observed genotype datasets from each context: the bundled maize panel, the HapMap Phase II subset distributed with SNPRelate for binary phenotypes, and the heterogeneous-stock mice distributed with BGLR for relationship matrices and prediction. The genotype input chapter shows how each dataset is prepared.
Three ways to read
- Learn a workflow: follow the worked chapters in order.
- Find a function: use the public function index.
- Inspect a method: follow the implementation overview, developer index, and equation index.
Validation against established software
Where a published tool already implements a method, we used it as an independent check of our implementation. Each comparison is reproducible from the repository.
| Reference software | What was checked | How |
|---|---|---|
| isqg 1.4 | The meiosis core behind crossing, selfing, and doubled haploids | Exact agreement with stored isqg outputs from the same random draws, required by the package tests |
| PLINK 1.9 | Composite and haplotypic LD pruning, variance-inflation pruning, haplotype blocks | Agreement with stored PLINK outputs in the package tests |
| AlphaSimR 2.1.0 | Genotype-by-environment phenotypes; coalescent founder presets; default crossover interference; residual draws and entry means | Exact agreement for genotype-by-environment phenotypes; distributional comparison for coalescent founders; scripts in dev/ |
Agreement on these methods does not make the packages interchangeable. Each chapter states which conventions follow another tool and where simplePHENOTYPES makes its own choices.
Version and scope
The examples are rendered against the package source in the same checkout as the book. The source revision and package version are recorded in the build record.
The book covers the modern API and keeps a dedicated chapter for the frozen create_phenotypes() interface (legacy interface). The original package publication is Fernandes and Lipka (2020); the current development interface includes methods added after that publication.