MAGIC-IRRI: Synthetic Plant Genetics Data for Trait Prediction

MAGIC populations are ideal for learning complex models due to their high genetic recombination, diversity, and large sample size. This synthetic dataset of 2000 observations was generated from a Bayesian network model developed for a talk on multiple trait prediction in plant genetics. The model and data were created by Marco Scutari, Phil Howell, David J Balding, and Ian Mackay.

Use Cases

• Predicting plant height (HT) based on other genetic traits as described in the model.
• Exploring Bayesian network structure learning for multiple trait modelling in genetics.
• Benchmarking machine learning algorithms on synthetic genetic recombination data.
• Studying the properties of Multiparent Advanced Generation Inter-Cross (MAGIC) populations for model training.

Strengths

• Data was generated from a model presented at the 5th International Conference on Quantitative Genetics (ICQG 2016).
• The description explicitly states the target variable is defined as 'HT' for prediction tasks.
• The underlying MAGIC populations are noted for high genetic recombination and diversity.

Limitations

• Row count is unknown for the original model data; the provided synthetic sample has 2000 rows.
• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

bnlearn R package and the Genetics Society of America.

Collection Method

Synthetic data generated (rbn) from a Bayesian network model developed for plant genetics.

Time Range

null

Freshness

unknown

Geography

null