GenoType-based Trio-binning (GT-Trio)

This is a Snakemake pipeline for creating de novo haplotype-resolved genome assemblies with trio-binning, using phased genotypes as parental input.

The pipeline consists of five steps: (1) parental sequence reconstruction, (2) parental read simulation (3) parental k-mer dictionary construction, (4) haplotype-resolved assembly and (5) reference-guided scaffolding.

Pipeline description

Step 1: Parental sequence reconstruction

Software: BCFtools (v1.18)

Parent-specific genome sequences are reconstructed from a high-quality reference genome (FASTA) and a dataset of phased parental genotypes (VCF) with BCFtools consensus. To account for phasing, two parent-specific sequences, representing either haplotypes are reconstructed using the --haplotype option.

Step 2: Parental read simulation

Software: SAMtools (v1.18)

To mimic Illumina short-read data used in conventional trio-binning, paired-end 150 bp reads are simulated from the parent-specific genome sequences with SAMtools wgsim. A user-defined number of read pairs -N is set to achieve the desired coverage of parental reads.

Step 3: Parental k-mer dictionary construction

Software: Yak (v0.1)

Parental k-mer dictionaries are created from simulated parental reads with Yak. An appropriate k-mer size has to be chosen to ensure a dictionary of unique paternal and maternal k-mers. See Koren et al. (2018) on how to choose an appropriate value of k for your species.

Step 4: Haplotype-resolved assembly

Software: Hifiasm (v0.24.0)

The paternal and maternal haplotypes of the trio offspring individuals are assembled from long-reads (ONT) using the trio-binning feature implemented in Hifiasm.

Step 5: Reference-guided scaffolding

Software: RagTag (v2.1.0)

Assembled contigs outputted from Hifiasm are combined into chromosome-level scaffolds with RagTag against the high-quality reference genome.

Usage

Environment setup

Make sure that Conda and Snakemake is installed.

Clone git repo

git clone https://github.com/theahettasch/GT-Trio.git
cd GT-Trio

Define input files and parameters

To run the pipeline with you own data, open the configuration file config/config.yaml and define any required input files and parameters (see config/README.md).

Run pipeline

When necessary dependencies (Conda and Snakemake) are available and any required input files and parameters have been defined in the configuration file, the GT-Trio pipeline can be run with one of the following commands from inside the GT-Trio directory:

#Run locally. Be aware that some of the tools implemented in the pipeline require certain amounts of memory and threads. 

snakemake \
  --use-conda \
  --cores 8 \ 
  --resources mem_mb=16000 

or

#Run on SLURM cluster. Required memory and threads is allready defined individually in each snakemake rule.

#!/usr/bin/env bash
#SBATCH --job-name=GT-Trio
#SBATCH --cpus-per-task=1
#SBATCH --mem=20G
#SBATCH --partition=partition_name
#SBATCH --output=logs/GT-Trio_%j.out
#SBATCH --error=logs/GT-Trio_%j.err

snakemake \
  --use-conda \
  --jobs 10 \
  --cluster "sbatch --mem={resources.mem_mb} --cpus-per-task={threads} --output={log[0]}" 

Output files

The pipeline outputs the following files:

parent_data/*aternal_sequence_hap1.fasta Reconstructed paternal/maternal sequence of haplotype 1

parent_data/*aternal_sequence_hap2.fasta Reconstructed paternal/maternal sequence of haplotype 2

parent_data/*aternal_reads.fq.gz Paternal/maternal simulated reads

parent_data/*aternal_count.yak Paternal/maternal k-mer dictionaries

hifiasm/offspring_assembly.hifiasm.asm* Output from Hifiasm. See Hifiasm output for details.

ragtag/ragtag.scaffold* Output from RagTag. See RagTag output for details.

assembly/offspring_assembly_contig_paternal.fa Offspring assembly of paternal haplotype (contigs)

assembly/offspring_assembly_contig_maternal.fa Offspring assembly of maternal haplotype (contigs)

assembly/offspring_assembly_chr_paternal.fasta Offspring assembly of paternal haplotype (chromosome scaffolds)

assembly/offspring_assembly_chr_maternal.fasta Offspring assembly of maternal haplotype (chromosome scaffolds)

Citing

I you use the pipeline, please site this paper: Haplotype assembly without parental sequencing: Genotype-based trio-binning (GT-Trio)