This repository is currently under development. Please use PsiPartition instead.
PsiPartitionPlus is a modified PsiPartition (xu-shi-jie/PsiPartition) (Xu & Onoda, 2024) with support for binary data and selectivity of whether users will apply ascertainment bias corrections (+ASC; Lewis, 2001) to partitions, in the IQ-TREE calculation phases, depending on the conditions of their data.
In my view, applying +ASC to a PsiPartition-determined partition in an alignment containing invariant sites that happens to have no invariant sites, which often occurs in the IQ-TREE calculation in PsiPartition, is not justified. It is because such a partition simply reflects the natural condition, rather than being an artificially altered version where invariant sites that should have originally existed are removed.
I thank Dr. Xu and Professor Onoda for developing the innovative software PsiPartition.
Hiroaki Sato
[email protected]
conda create -n PsiPartitionPlus
conda activate PsiPartitionPlus
pip install -r requirements.txtTo enable the Bayesian optimization, you need to register an account on Weights & Biases.
python PsiPartitionPlus_wandb.py --msa <MSA File> --format <fasta or phylip> --alphabet <dna, aa, or bin> --asc <no or yes> --max_partitions <max_partitions> --n_iter <number of iterations>--asc specifies whether +ASC should be applied to all partitions or not at all in the IQ-TREE calculation phases.
If you used PsiPartitionPlus in your work, please cite the original PsiPartition paper (Xu & Onoda, 2024) and the Zenodo repository of PsiPartitionPlus (click Cite this repository).