Third generation sequence assembly

A Snakemake pipeline for third-generation sequence assembly, designed to automate the processing and assembly of third-generation sequencing data through quality control, assembly, polishing, and post-processing.

Basic Information

Item	Details
Name	Third generation sequence assembly
Date	2023-04-24
Description	Snakemake pipeline for automated assembly of third-generation sequencing data, including quality control, de novo assembly, multi-round polishing, and genome optimization.
Author	Tong Xu

Dependencies

The pipeline relies on the following tools (specific versions recommended):

• fastp = 0.23.2 (short-read quality control)
• flye = 2.9.2 (third-generation sequence de novo assembly)
• minimap2 = 2.24 (sequence alignment)
• samtools = 1.17 (BAM file processing)
• pilon = 1.24 (genome polishing)
• prodigal (ORF prediction for post-processing)

Directory Structure

Input Directories

Input data must follow this structure (relative to base_dir in the script):

• SG_reads/{sample}/: Short-read (second-generation) data, named as {sample}_1.fq (R1) and {sample}_2.fq (R2).
• TG_reads/{sample}/: Third-generation sequencing data, named as {sample}.filtered_reads.fq.

Output Directories

Automatically generated output directories:

• fastp_output/: Quality-controlled short reads and fastp reports (JSON/HTML).
• flye_output/: Initial assembly results from Flye.
• minimap2_flye_output/: Alignment files (BAM) generated by minimap2.
• pilon_flye_output/: Polished genomes after each Pilon round.
• TG_unfixed_genome/: Renamed contigs before final optimization.
• prodigal_output/: ORF prediction results (GFF format) from Prodigal.
• TG_genome/: Final optimized genome assemblies.

Pipeline Workflow

The pipeline runs in the following sequential steps via Snakemake rules:

1. Quality Control with fastp Trims low-quality bases, adapters, and filters short reads from short-read data. Outputs cleaned reads and quality reports (JSON/HTML).
2. De novo Assembly with Flye Assembles third-generation reads into initial contigs. Key parameters:
   • Genome size: ~4.5m (optimized for Phaeobacter gallaeciensis; adjust based on target species).
   • Coverage limit: 300x (prevents assembly failure due to excessive depth).
3. Multi-round Genome Polishing
   • minimap2: Aligns cleaned short reads to the assembled genome, generating sorted BAM files.
   • Pilon (3 rounds): Polishes the genome using short-read alignments to correct SNPs and indels, with a minimum depth threshold of 10 for reliable corrections.
4. Contig Renaming Renames contigs to standardized identifiers (e.g., {sample}_chromosome_1, {sample}_plasmid_1) based on assembly metrics (length, coverage, circularity from assembly_info.txt).
5. ORF Prediction with Prodigal Predicts open reading frames (ORFs) from renamed contigs to guide genome optimization.
6. Circular Contig Fixing Adjusts circular contigs to repair potential broken genes by reorienting sequences based on ORF positions (moves non-ORF regions to the end of contigs).

Usage

1. Prepare Input Data: Organize short-read and third-generation data into the required directory structure.

2. Configure Paths: Ensure base_dir (root directory) and pilon_jar (path to Pilon JAR file) in the script are correctly set.

3. Run the Pipeline:

   Execute with Snakemake, specifying the number of cores:

   bash

   snakemake --cores [number_of_cores]

Notes

• Genome Size Adjustment: Modify the genomeSize parameter in the Flye rule for non-Phaeobacter gallaeciensis species.
• Polishing Stringency: Pilon uses --mindepth 10 to filter low-coverage variants; adjust based on data quality.
• Circular Contigs: The final step fixes potential gene breaks in circular contigs, improving genome completeness for downstream analysis.