📚 Snakemake pipeline: genome-reference-builder

This README provides detailed instructions for using the Genome Reference Builder, a Snakemake pipeline that automates the process of genome indexing for organisms such as human and mouse. It integrates tools like STAR, RSEM, and Cell Ranger, streamlining the genome preparation process for downstream analysis.

🌟 Introduction

The Genome Reference Builder pipeline simplifies and automates the creation of genome indices, which are essential for various genomic analyses. This pipeline integrates multiple bioinformatics tools to ensure a smooth and efficient genome indexing process. The pipeline is built around GENCODE annotation data and is currently set to use October 2024 versions:

• Human: GENCODE Release 47 (GRCh38.p14)
• Mouse: GENCODE Release M36 (GRCm39)

⚠️ Note: These versions will be used for the foreseeable future in all my personal analyses.

🔄 Pipeline overview

The following flowchart shows the main steps of the Genome Reference Builder pipeline and their interrelationships:

1. Download reference data: Fetch the latest GENCODE annotation and genome sequence files.
2. Filter GTF files: Apply filtering criteria to GTF files to retain only relevant gene and transcript types.
3. Generate genome indices: Create genome indices using various bioinformatics tools such as STAR, HISAT2, Kallisto, Salmon, RSEM, and Cell Ranger.
4. Validate outputs: Ensure the generated indices are correct and consistent with the input data.

Each step is interconnected, ensuring a streamlined and automated process for preparing genome references.

📋 Prerequisites

Before running the pipeline, ensure that you have installed the following prerequisites:

• Mamba (Mamba or Micromamba distribution)
• Snakemake (version 8.24.0 or higher)

Additionally, the pipeline relies on specific versions of bioinformatics software:

Software	Version	Reference (PMID)
HISAT2	2.2.1	31375807
STAR	2.7.11b	23104886
Kallisto	0.51.1	27043002
Salmon	1.10.3	28263959
RSEM	1.3.3	21816040
Cell Ranger	8.0.1	28091601

⚠️ Note: To ensure consistency between Bulk RNA-seq and Single-cell RNA-seq data analyses, the genome indexes constructed by this pipeline are based on filtered GTF files following the 10x Genomics filtering scheme. When building the reference genome, the following software utilize the filtered GTF files as input: RSEM, STAR, and Cell Ranger.

The Genome Reference Builder pipeline filters GTF files based on specific gene types and transcript tags to ensure high-quality annotations. Below are the detailed criteria used during the filtering process:

Category	Criteria	Description
Gene Types	protein_coding, protein_coding_LoF, lncRNA, IG_C_gene, IG_D_gene, IG_J_gene, IG_LV_gene, IG_V_gene, IG_V_pseudogene, IG_J_pseudogene, IG_C_pseudogene, TR_C_gene, TR_D_gene, TR_J_gene, TR_V_gene, TR_V_pseudogene, TR_J_pseudogene	Only genes belonging to these types are retained.
Transcript Types	protein_coding, protein_coding_LoF, lncRNA, IG_C_gene, IG_D_gene, IG_J_gene, IG_LV_gene, IG_V_gene, IG_V_pseudogene, IG_J_pseudogene, IG_C_pseudogene, TR_C_gene, TR_D_gene, TR_J_gene, TR_V_gene, TR_V_pseudogene, TR_J_pseudogene	Only transcripts of these types are retained.
Excluded Tags	readthrough_transcript	Transcripts tagged as readthrough are excluded.
Version Suffix	Removed from gene_id, transcript_id, and exon_id	Ensures consistency by removing version numbers from IDs.

To evaluate the effectiveness of the filtering process, we present a summary of gene counts before and after filtering for both human and mouse GTF files (October 2024 version).

• Human

  GTF	Total	Protein-coding	lncRNA
  Unfiltered	77,307	20,089	35,045
  Filtered	54,907	19,446	34,818

• Mouse

  GTF	Total	Protein-coding	lncRNA
  Unfiltered	77,981	21,729	32,947
  Filtered	55,095	21,530	32,873

🛠️ Installation

1. Install Snakemake using Mamba 🐍:

   mamba install -c bioconda snakemake

2. Clone the pipeline repository from GitHub 🐙:

   git clone git@github.com:catplot/genome-reference-builder.git

3. Navigate to the pipeline directory 📂:

   cd genome-reference-builder

⚙️ Configuration

The pipeline uses a configuration file (config.yaml) to specify the reference genomes and indexing tools. This file is included in the repository and is already set up for the October 2024 GENCODE versions of human and mouse genomes. No further modifications to the configuration are required unless you want to adjust specific paths or parameters.

🚀 Usage

There are two ways to run the Genome Reference Builder pipeline: directly or via a SLURM job scheduler.

Option 1: Direct execution

You can directly execute the pipeline using the following command:

snakemake \
--configfile config/config.yaml \
--cores 16 \
--use-conda \
--rerun-incomplete \
--keep-going

Option 2: Using SLURM Job Scheduler

To run the pipeline on a SLURM cluster, use the following command to submit jobs to SLURM:

snakemake \
--configfile config/config.yaml \
--cores 16 \
--use-conda \
--rerun-incomplete \
--keep-going \
--cluster "sbatch --job-name=genome-indexing --mem={resources.mem_mb} --cpus-per-task={threads} --time={resources.runtime} --output=logs/snakemake_%j.out --error=logs/snakemake_%j.err" \
--jobs 10

📂 Output

The pipeline generates indexed genome files essential for downstream genomic analysis. The output files are organized in the directory specified in the configuration file, following this structure:

/datapool/reference_genomes/
└── gencode/
├── human/
│ └── 47/
│ ├── GRCh38.primary_assembly.genome.fa
│ ├── GRCh38.primary_assembly.genome.fa.fai
│ ├── GRCh38.v47.transcripts.fa
│ ├── GRCh38.v47.primary_assembly.annotation.gtf
│ ├── index/
│ │ ├── star/
│ │ ├── hisat2/
│ │ ├── kallisto/
│ │ ├── salmon/
│ │ ├── rsem/
│ │ └── cellranger/
└── mouse/
└── M36/
├── GRCm39.primary_assembly.genome.fa
├── GRCm39.primary_assembly.genome.fa.fai
├── GRCm39.vM36.transcripts.fa
├── GRCm39.vM36.primary_assembly.annotation.gtf
├── index/
│ ├── star/
│ ├── hisat2/
│ ├── kallisto/
│ ├── salmon/
│ ├── rsem/
│ └── cellranger/

🐛 Troubleshooting

• Ensure all dependencies are installed correctly and the paths in your config file are properly set.
• For issues related to Snakemake execution, refer to the Snakemake documentation 📖.

📞 Further help

For more details or help with specific tools, consult the documentation for each bioinformatics tool used in this pipeline. You can also open an issue on the repository's GitHub page for additional assistance.