Introduction to tstrait

Welcome to tstrait, a quantitative trait simulator based on succinct tree sequences. Tstrait supports simulation of quantitative traits under various additive models, with effect sizes taken from specified probability distribution including, but not limited to, the normal, t-, or Gamma distributions. Details of supported distributions are in (Effect Size Distribution). It also supports multi-trait simulation under pleiotropy (see Multi-trait simulation for details). For ease of use, the output of tstrait is given as a pandas.DataFrame.

Advantages of tstrait

Tstrait is built on top of tskit, and uses tree sequence data as an input. Tree sequences are designed to efficiently store and process millions of DNA sequences. As a result, tstrait can simulate quantitative traits substantially faster than working with the genotype matrix or other traditional data structures.

Quantitative trait simulation in tstrait is transparent, and users can control each step in the simulation. Thus, it is possible for the users to simulate their own environmental noise on top of simulated genetic values, or even use their own defined effect sizes and causal sites. The tree sequence data structure is widely used in various population genetic simulation packages, including SLiM, msprime, and stdpopsim; it is therefore easy for users of these packages to add quantitative traits to their results using tstrait.

Tree Sequence resources

To learn more about tree sequences:

• The tskit website provides learning materials explaining what tree sequences are, and includes tutorials, publications and videos.

• The PySLiM manual explains how forward genetic simulation can be create tree sequences.

• The msprime manual details an efficient backward-time genetic simulator that outputs tree sequences.

• The tskit tutorials explain how to analyze succinct tree sequences by using tskit.