Mandala Compared With Common R Breeding and Genomic Prediction Packages
The ratings summarize practical strengths for crop breeding workflows, including trial analysis, genomic prediction, usability, speed, memory behavior, and end-to-end analysis support. Ratings are qualitative and intended for positioning rather than formal benchmarking.
Rating scale:
★★★ strong,
★★ moderate/usable,
★ limited or not the primary purpose.
| Criterion | Mandala | sommer | rrBLUP | lme/nlme/lme4 | BGLR |
|---|---|---|---|---|---|
| General breeding experiment analysis | ★★★ | ★★ | ★ | ★★ | ★ |
| Single-site trial models | ★★★ | ★★ | ★ | ★★ | ★ |
| MET analysis | ★★★ | ★★ | ★ | ★ | ★★ |
| Spatial / row-column analysis | ★★★ | ★★ | ★ | ★★ | ★ |
| Advanced variance structures | ★★ | ★★ | ★ | ★★ | ★★ |
| Genomic prediction | ★★★ | ★★★ | ★★★ | ★ | ★★★ |
| Genomic + field-design integration | ★★★ | ★★ | ★ | ★ | ★★ |
| Multi-trait / multivariate capacity | ★ | ★★★ | ★ | ★★ | ★★★ |
| Speed for routine breeding workflows | ★★★ | ★★ | ★★★ | ★★ | ★★ |
| Memory-safe large model behavior | ★★★ | ★★ | ★★ | ★★ | ★★ |
| Efficiency optimization potential | ★★★ | ★★ | ★★ | ★ | ★★ |
| Breeder friendliness | ★★★ | ★★ | ★★ | ★ | ★★ |
| End-to-end breeding pipeline | ★★★ | ★★ | ★ | ★ | ★★ |
| Prediction API and SEs | ★★★ | ★★ | ★ | ★ | ★★ |
| Teaching / reproducible workflow | ★★★ | ★★ | ★★ | ★★ | ★★ |
Interpretation: Mandala is strongest as a fast, breeder-friendly, end-to-end framework for crop trial analysis, MET modeling, genomic prediction integration, prediction SEs, fixed-effect tests, diagnostics, and reproducible workflows. Its multivariate functionality should be treated as early-stage relative to packages with mature multi-trait engines.