Choose a genomic analysis path

Getting Started

Start with your breeding question and choose a path through genotype preparation, genomic analysis, pedigree or ancestry assessment, and mating decisions.

Author

Alexander M. Sandercock

Published

July 13, 2026

Modified

July 14, 2026

Keywords

genomics, workflow, tool selection, analysis planning, ploidy, GWAS, genomic prediction, pedigree, parentage, ancestry, mate allocation

You do not need to choose software before you understand the breeding question. Start with the decision you want to make, then check whether you have the data needed to support it.

The useful question is not “Which package should I use?” It is “What do I need to learn or decide, and what information do I already have?”

This guide connects six Breeding Insight tools. You may use one tool for a bounded task or move through several as genotype data become breeding decisions. Not every program needs every step.

Start with the decision

Before opening an application or writing R code, state the immediate question in one sentence. For example:

Are these genotype data ready for analysis?
What is the ploidy or large-scale copy number of each sample?
Are the samples genetically structured?
Which markers are associated with a measured trait?
How well can these data predict performance in the intended selection candidates?
Does the recorded pedigree agree with the marker data?
Which candidate parents balance expected gain and relatedness?

Each question requires different inputs and produces a different kind of evidence. A principal component analysis cannot validate a recorded parent, and a parentage result does not provide the breeding values needed to optimize crosses.

Check what you have

Use the following checklist before choosing a path.

Genotype evidence: Do you have read counts, a genotype-called VCF, another marker matrix, or only raw sequence files?
Sample identity: Do sample IDs match across genotype, phenotype, pedigree, and candidate files?
Ploidy: Is the expected ploidy known, and is it plausible for every sample?
Genome information: Do markers have chromosome and position information when the analysis requires it?
Phenotypes: Do you have a measured trait for association analysis, prediction, or estimated breeding values from a suitable evaluation?
Relationships: Do you have a pedigree, candidate-parent pool, reference populations, or other relationship information required by the question?

If one of these is missing, the next step may be data preparation rather than analysis.

Follow the path that matches the question

The diagram is a decision map, not a required sequence. Some projects enter at pedigree assessment or mate allocation because their genotype processing and genetic evaluation are already complete.

%%{init: {"flowchart": {"rankSpacing": 34, "nodeSpacing": 46, "curve": "basis"}}}%%
flowchart TB
  accTitle: Breeding Insight genomic analysis decision path
  accDescr: A breeding question leads to checks of available inputs and ploidy. Genotype preparation can use BIGr or BIGapp, ploidy and copy-number assessment can use Qploidy2, genomic analysis can use BIGr or BIGapp, pedigree and ancestry assessment can use BIGpopA or Familia, and candidates with pedigree and breeding values can continue to mate allocation with AlloMate.

  A([Breeding question]) --> B{{Check inputs,<br/>IDs, and ploidy}}
  B --> C[/Prepare and inspect<br/>genotype data/]
  B --> D[/Estimate ploidy or<br/>large-scale copy number/]
  C --> E([BIGr or BIGapp])
  D --> F([Qploidy2])
  E --> G{{What is the next<br/>breeding question?}}
  F --> G
  G --> H[/Structure, GWAS, or<br/>genomic prediction/]
  G --> I[/Pedigree, parentage,<br/>or ancestry/]
  H --> J([BIGr or BIGapp])
  I --> K([BIGpopA or Familia])
  J --> L{{Candidates, pedigree,<br/>and breeding values ready?}}
  K --> L
  L -->|yes| M([AlloMate])
  L -->|not yet| N([Prepare the missing<br/>selection inputs])

  classDef question fill:#EAF5F7,stroke:#2A6576,stroke-width:2px,color:#173D47
  classDef task fill:#FFF7E8,stroke:#C98B2E,stroke-width:2px,color:#674615
  classDef tool fill:#EEF8F0,stroke:#319B42,stroke-width:2px,color:#1E5D28
  classDef check fill:#F3EFF9,stroke:#765C9E,stroke-width:2px,color:#493566

  class A question
  class C,D,H,I task
  class E,F,J,K,M,N tool
  class B,G,L check

Choose tools from the question and required inputs, not from package names alone.

Choose the immediate task

Prepare or inspect marker data

Start with BIGr when you want a scripted R workflow that can be saved, reviewed, and adapted. Start with BIGapp when the supported analysis is available through its graphical interface and you prefer to work without command-line inputs.

Estimate ploidy or copy number

Use Qploidy2 when the question concerns individual ploidy, aneuploidy, or large-scale copy-number variation. Its workflow uses marker information such as allele ratios and depth or intensity across the genome.

Explore structure or map a trait

Use BIGr for a scripted workflow or BIGapp for its graphical PCA, DAPC, GWASpoly, and genomic-prediction workflows. Phenotypes and marker positions are needed only for analyses that use them, but genotype quality and sample identity matter throughout.

Check pedigrees or assign parents

Use BIGpopA when you want an R workflow for pedigree quality control, marker-based parentage assignment, or breed or line composition. Use Familia when you want a graphical interface to BIGpopA workflows and ancestry analyses.

Assess ancestry

Use Familia for interactive supervised breed or line composition and unsupervised ancestry inference. Use BIGpopA when the supervised composition workflow needs to be run and recorded in R.

Choose contributions and crosses

Use AlloMate after the candidate list, pedigree, and estimated breeding values are ready. It combines pedigree-based kinship, trait weights, contribution constraints, and mating rules to support selection and crossing decisions.

Compare the starting points

Immediate question	Have ready	Start with	What it contributes
How do I process or inspect genotype data?	Read counts, VCF, or another supported genotype input; sample IDs; expected ploidy	BIGr or BIGapp	Scripted or graphical genotype processing, quality summaries, and supported downstream analyses
Are ploidy or large chromosome-scale changes uncertain?	Genotyping-platform allele ratios plus depth or intensity and marker positions appropriate for the workflow	Qploidy2	Ploidy, aneuploidy, and large-scale copy-number estimates
Are samples structured, associated with a trait, or useful for prediction?	Quality-controlled genotypes; phenotypes and marker positions when required	BIGr or BIGapp	Population summaries, trait-association evidence, or validated genomic predictions supported by the selected workflow
Does the pedigree agree with marker data, or who are the likely parents?	Pedigree or candidate-parent pool plus compatible SNP genotypes	BIGpopA or Familia	Pedigree checks, parentage assignments, and related diagnostics
What is the genomic ancestry or breed or line composition?	Target genotypes and, for supervised analysis, suitable reference populations	BIGpopA or Familia	Supervised composition estimates or, in Familia, unsupervised ancestry inference
Which candidates should contribute, and how should they be mated?	Candidate IDs, pedigree, required candidate metadata, and estimated breeding values	AlloMate	Kinship summaries, weighted selection information, contribution results, and mating plans

The two paired choices are mainly about how you want to work and which analysis is required:

BIGr and BIGapp: BIGr exposes R functions for scripted and adaptable workflows. BIGapp exposes supported genomic analyses through a graphical Shiny interface and uses BIGr within its software ecosystem.
BIGpopA and Familia: BIGpopA provides the underlying R analyses for pedigree, parentage, and supervised composition tasks. Familia provides an interactive interface to those workflows and adds unsupervised ancestry analysis.

Three example routes

Route 1: Explore diversity and map a measured trait

Suppose you have a genotype-called VCF, sample metadata, a phenotype, and SNP positions. You want to inspect data quality, explore structure, and test marker-trait associations without writing R code.

Install and run BIGapp locally. The public demo is for viewing the interface; use a local installation for the workshop and your own analysis.
Confirm sample IDs and ploidy, then inspect and filter the genotype data.
Run PCA and use the structure results to inform the GWAS model.
Run GWASpoly and interpret the Manhattan, QQ, and linkage-disequilibrium results together.

Start the BIGapp workshop

Route 2: Turn targeted read counts into analysis-ready genotypes

Suppose you received a processed MADC file from targeted sequencing and want a reproducible VCF workflow.

Confirm that the MADC has fixed allele IDs and that sample names match the project records.
Use BIGr to recover biallelic read-count evidence, call genotypes with the correct ploidy, and evaluate the model diagnostics.
Mask uncertain cells and filter markers and samples using evidence from the dataset.
Continue in BIGr or load the genotype-called VCF into BIGapp for a supported graphical analysis.

Understand MADC files Follow the MADC-to-VCF workflow

Route 3: Validate records before choosing crosses

Suppose you have SNP genotypes, a recorded pedigree, a candidate-parent pool, and estimated breeding values.

Use Familia for an interactive workflow or BIGpopA for a scripted workflow to clean the pedigree, evaluate Mendelian errors, and assign parentage where the data support it.
Review unresolved samples and update records only when the evidence and breeding-program context justify the change.
Prepare the candidate list, pedigree, required candidate metadata, and breeding values for AlloMate.
Use AlloMate to evaluate kinship, trait weights, contribution constraints, and proposed matings.

Pedigree validation and mate allocation answer different questions. Do not treat a plausible parentage assignment as a breeding value or a mating recommendation.

Understand validation and assignment Run the BIGpopA tutorial

Decisions the software cannot make for you

The software can calculate, summarize, and visualize the evidence you provide. It cannot determine whether the biological design answers the breeding question.

Before acting on a result, ask:

Does the sample set represent the population where the conclusion will be used?
Were ploidy, genotype encoding, and uncertainty handled correctly?
Are missingness, batch effects, and population structure changing the result?
Does the validation design represent the future candidates you intend to predict?
Are pedigree, reference-population, phenotype, and candidate IDs aligned across files?
Can the result be reproduced from recorded inputs, software versions, and settings?

If the answer is uncertain, return to data checks or compare reasonable analysis choices before moving to the next tool.

Continue learning

Get started with genomic data in R for a worked BIGr and VCF introduction.
What is allele dosage? before interpreting diploid or polyploid genotype states.
Choosing a missing-data threshold before copying a filtering rule from another dataset.
SNP filtering in polyploids for depth, dosage uncertainty, and suspected duplicated-locus checks.
What does cross-validation accuracy mean? before using a genomic-prediction result for selection.
Pedigree validation or parentage assignment? before interpreting a pedigree conflict or candidate-parent result.
Validate pedigrees and assign parentage with BIGpopA for a worked R example.
Estimate copy number in alfalfa with Qploidy2 for a complete Qploidy2 example.