# Data-specific substitution models improve protein-based phylogenetics (Brazao et al.)

## Data-specific amino-acid substitution models estimated from the simulated sequence alignments with 400, 1500, and 8000 sites.

## 1. Codeml-estimated models were generated using the Codeml configure file (codeml_template.ctl file with values and parameters in this folder).

## 2. IQ-TREE-estimated models were estimated using the following command: 'iqtree -s alignment_file -te simulation_tree -blfix -m GTR20+G4+FO'

## 3. FastMG-estimated models were calculated using the following command: 'perl estimate.pl -a alignmentFolder -s gcpREV.dat -o data_specific_model'
## The tree search step of the estimate.pl (line 188) was changed to enable to fix the simulation tree (estimate.pl in this folder).

## 4. P4-BI-estimated models were generated using the custom scripts (deposited in: https://github.com/joaobrazao/Data-specific-substitution-models-improve-protein-based-phylogenetics):
## mcmc_for_estimating_rates.py (for estimating model parameters from a MCMC run) and AminoacidsModelFormatConvertor.py (convert the parameters to a paml model format).

## 5. P4-ML-estimated models were generated using the custom script https://github.com/joaobrazao/Data-specific-substitution-models-improve-protein-based-phylogenetics/calculate_model_parameters_ML_P4.py

## Folders list:
├── Codeml_models
├── FastMG_models
├── IQTREE_models
├── P4_BI_models
└── P4_ML_models

