Introduction

This page offers detailed introduction of HybSuite. Feel free to explore!

• Pipeline overview
• Features
• Advantages

🧬 Pipeline overview

HybSuite performs end-to-end hybrid capture (Hyb-Seq) phylogenomic analysis from raw reads (Hyb-Seq preferred; compatible with RNA-seq, WGS, and genome skimming data) to phylogenetic trees.

The full pipeline is composed of 4 stages:

• Stage 1: NGS dataset construction

  • (1) Optionally download public raw reads from NCBI (via SRA Toolkit );
  • (2) Optionally integrate user-provided raw reads (if provided);
  • (3) Raw reads trimming (via Trimmomatic);
    
    

• Stage 2: Data assembly and paralog retrieval

  • (1) Target loci assembly and putative paralogs retrieval (via HybPiper)
  • (2) Integrate pre-assembled sequences (if provided);
  • (3) Filter putative paralogs;
  • (4) Plot recovery heatmap and paralog heatmap of original and filtered sequences;
    
    

• Stage 3: Paralog handling

  • Optionally execute seven paralogs-handling methods (HRS, RLWP, LS, MO, MI, RT, 1to1; see our Tutorial and generate filtered alignments for downstream analysis:
    • HRS:
      (1) Retrieve seqeunces via command hybpiper retrieve_sequences in HybPiper;
      (2) Integrate pre-assembled sequences (if provided);
      (3) Filter sequences by length to remove potential mis-assembled seqeunces;
      (4) Mutiple sequences aligning (via MAFFT) and trimming (via trimAl or HMMCleaner);
      (5) Filter trimmed alignments to generate final alignments.
    • RLWP:
      (1) Retrieve seqeunces via hybpiper retrieve_sequences via HybPiper;
      (2) Integrate pre-assembled sequences (if provided);
      (3) Filter sequences by length to remove potential mis-assembled seqeunces;
      (4) Remove loci with putative paralogs masked in more than samples;
      (5) Mutiple sequences aligning (via MAFFT) and trimming (via trimAl or HMMCleaner);
      (6) Filter trimmed alignments to generate final alignments.
    • PhyloPypruner pipeline (LS, MI, MO, RT, 1to1):
      (1) Mutiple sequences aligning (via MAFFT) and trimming (via trimAl or HMMCleaner) for all putative paralogs;
      (2) Gene trees inference of all putative paralogs;
      (3) Obtain orthogroup alignments using tree-based orthology inference algorithms (via PhyloPypruner);
      (4) Realign (via MAFFT) and trim (via trimAl or HMMCleaner) the orthogroup alignments;
      (5) Filter trimmed orthogroup alignments to generate final alignments.
    • ParaGone pipeline (MI, MO, RT, 1to1):
      (1) Use the directory cantaining all putative paralogs generated in stage 2 as input;
      (2) Obtain orthogroup alignments using tree-based orthology inference algorithms via ParaGone;
      (3) Filter trimmed orthogroup alignments to generate final alignments.
      
      

• Stage 4: Species tree inference

  • Multiple species tree inference methods available:
    • Concatenation-based approach: IQ-TREE, RAxML, or RAxML-NG;
    • Coalescent-based approach: ASTRAL-IV or wASTRAL;
    • Multi-copy genes aware coalescent-based approach: ASTRAL-pro3.

✨ Features

🔄 Transparent: Full workflow visibility with real-time progress logging at each step
📝 Reproducible: Automatically archives exact software commands & parameters for every run
🧩 Modular: Execute individual stages or complete pipeline in one command
⚡ Flexible: 7 paralog handling methods & 5+ species tree inference options
🚀 Scalable: Built-in parallelization for large-scale phylogenomic datasets

🏆 Advantages

1. End-to-end pipeline from reads to trees

• Processes data from raw reads to phylogenetic trees with single-command workflows
• Supports both full pipeline execution and modular stage-specific operations
• Minimizes manual intervention while maintaining flexibility

2. Unique functionality of integrating pre-assembled sequences

• Allows for integrating pre-assembled loci sequences into the working dataset. (click here to grasp skills)

3. Customizable sequences filtering strategies

• Dual filtering strategies for both loci and samples
• Configurable thresholds for read depth, missing data, and sequence quality
• Enables dataset optimization for different study goals

4. Advanced paralog-handling methods

• Implements 7 distinct methods for paralog detection and processing
• Includes both similarity-based and topology-based approaches
• Improves orthology assessment accuracy

5. Multi-method Phylogenetic tree inference

• Integrated softwares for concatenation-based methods: IQ-TREE, RAxML, and RAxML-NG
• Integrated softwares for coalescent-based methods: ASTRAL-III or wASTRAL

6. Integrated visualization tools

• plot_paralog_heatmap.py (click here to grasp skills);
• plot_recovery_heatmap.py (click here to grasp skills)
• modified_phypartspiecharts.py (click here to grasp skills)

7. High-Performance Computing

• Parallel processing across samples and loci (option -process), which can significantly improve computational efficiency.