FastMultRec: Fast Multiple gene trees Reconciliation program

Compiling

To compile the FastMultRec program, you can use the cmake and make approach as follows:

mkdir build
cd build
cmake ..
make

This will produce the FastMultRec executable.

Alternatively, to compile manually, use this command:

g++ -std=c++11 -O3 main.cpp define.h genespeciestreeutil.h hashtable.h newicklex.h node.h SegmentalReconciler.h treeinfo.h treeiterator.h util.h ReconciliationTester.h genespeciestreeutil.cpp newicklex.cpp node.cpp SegmentalReconciler.cpp treeinfo.cpp treeiterator.cpp ReconciliationTester.cpp -o FastMultRec

(Note: Additional optimization flags can be added if required.)

---

Usage

To run the program, use the following example command:

./FastMultRec -d 5 -l 1 -gf ../data/gene_trees.txt -sf ../data/s_tree.newick -spsep "_" -spindex 0 -o output.txt -al FastMultRec

• -al specifies the reconciliation algorithm. Options:
  • simphy: Use the original mapping from SimPhy.
  • LCA: Use the LCA mapping.
  • Any other value: Use the FastMultRec remapping algorithm.

Required Arguments:

• At least one of -g or -gf must be provided, and at least one of -s or -sf must be provided.
  • -g [g1;g2;...;gk]: Gene trees in Newick format, separated by ;.
  • -gf [file]: File containing gene trees in Newick format, separated by ;.
  • -s [newick]: Species tree in Newick format.
  • -sf [file]: File containing the species tree in Newick format.
  • -al [LCA, simphy, FastMultRec]: Algorithm to use for reconciliation.

Optional Arguments:

• --help: Display help message.
• -d [double]: Cost for one height of duplication. Default: 3.
• -l [double]: Cost for one loss. Default: 1.
• -h [int]: Maximum allowed duplication sum-of-heights. Default: 20.
• -o [file]: Output file. Default: output to console.
• -spsep [string]: Separator between gene and species in gene names. Default: __.
• -spindex [int]: Position of the species in gene names after splitting by the separator. Default: 0.
• --test: Run a series of unit tests, including fixed examples and random trees for more complex testing.

---

SimPhy

https://github.com/adamallo/SimPhy/wiki/Manual

Simulating with WGD and Applying Fractionation

Use the script supplementaries/simulate_1WGD_simphy.sh to insert a WGD event into a SimPhy simulation and apply fractionation. This script makes use of the following Python scripts:

• map_gene_trees_oneWGD.py
• apply_losses_on_simphy.py
• post-order-labeling.py

Note: map_gene_trees_oneWGD.py may need to be modified to handle different numbers of WGDs based on filenames in SimPhy simulation directories.

Analyzing a SimPhy Simulation Directory

Use the script supplementaries/analyze_sims_parallel.sh to calculate the path distances between different mappings (e.g., between LCA and SimPhy or between LCA and FastMultRec with varying duplication costs). This script relies on compare_mapping.py.

Calculating Recall and Precision

To calculate recall and precision for duplications, follow these steps:

1. Run supplementaries/dups-in-gis_run.sh to calculate the number of gene trees supporting duplications at each species (uses dups-in-gis.py).
2. Call supplementaries/DUPs-recall-prec_run.sh to calculate recall and precision (uses DUPs-recall-prec-V2.py).

Applying NNIs to SimPhy Simulations

Use the script supplementaries/apply_NNIs.sh to apply NNIs (Nearest Neighbor Interchanges) to SimPhy simulations. You can set the number of NNIs applied to each gene tree via the k parameter in the script. This script uses apply_NNIs.py.

Stochastic Version

To run the stochastic version of FastMultRec, use the following command:

./FastMultRec -d 5 -l 1 -gf ../data/gene_trees.txt -sf ../data/s_tree.newick -spsep "_" -spindex 0 -o output.txt -al stochastic -tmp 1 -stoloops 2000

Parameters:

• -d [double]: Cost for one height of duplication.
• -l [double]: Cost for one loss.
• -gf [file]: File containing gene trees in Newick format, separated by ;.
• -sf [file]: File containing the species tree in Newick format.
• -spsep [string]: Separator between gene and species in gene names (default: __).
• -spindex [int]: Position of the species in gene names after splitting by the separator (default: 0).
• -o [file]: Output file.
• -al [stochastic]: Specifies the use of the stochastic reconciliation algorithm.
• -tmp [double]: Specifies the temperature value for the stochastic algorithm.
• -stoloops [int]: Number of loops after which the algorithm stops if no improvement is detected.

🧪 Useful Scripts for Gene Tree Simulation and Analysis

This repository contains all supporting scripts used in our experiments, located under the supplementaries/Scripts/ directory.

---

📁 Directory Structure

simulate_gene_trees/

This directory includes all the scripts required to simulate species trees and gene trees using the SimPhy framework.

🔧 Main Script: all_sim.py

This script orchestrates the full simulation pipeline.

Features:

• Allows specifying the number of Whole Genome Duplications (WGD) to simulate:
  • 0: no WGD
  • 1: single WGD
  • 2: double WGD
• Automatically applies fractionation (gene loss) on the duplicated gene trees.
• Generates output compatible with downstream phylogenomic reconciliation analysis.

generate_reconstructed_gene_trees/

This section covers how to reconstruct gene trees from simulated sequences.

🧬 Step 1: Convert gene trees to sequences

Use the call_convert_genetrees_to_seq.py script, which wraps around Seq-Gen, to simulate sequence evolution from gene trees.

🧬 Step 2: Reconstruct gene trees from sequences

Use the run_iqtree.sh script to reconstruct gene trees using the IQ-TREE tool.

clean_constructed_gene_trees/

After gene tree reconstruction, some gene trees may contain low-support or erroneous branches. To clean them:

🧹 Script: clean_genetrees_with_ecceTera.sh

This script uses the ecceTERA tool to clean the reconstructed gene trees by reconciling them with a species tree and resolving low-confidence branches.

generate_FastmultRec_LCA_MetaEC_results/

You can use the analyze_sims_parallel_greedy_down.sh or run_metaec.py scripts to generate results from FastMultRec, LCA, and MetaEC methods.
You can specify whether to run them on the true gene trees or on the constructed gene trees.

---

Recall_Precision_Evaluation/

To evaluate recall and precision:

Step 1: Generate CSVs for individual simulations

Run the script:

bash DUPs-recall-prec_run.sh

This generates CSV files with recall and precision values for each simulation. You can modify the script to adjust the threshold used for evaluation.

Step 2: Compute average recall/precision

Run:

python dups-TP-avg.py

This aggregates the results across simulations and computes averages.

Step 3: Generate precision-recall plots

Use:

python generate_recall_precision_plots_comb.py

This produces plots comparing FastMultRec, LCA, and MetaEC across simulations.