LEMON

It is a software takes use of existing shotgun NGS datasets to detect HGT breakpoints, identify the transferred genome segments, and reconstructs the inserted local strain.

Table of Contents

1. Installation
2. LEMON usage
3. Example workflow

   Installation

   Requirements

   • Softwares
   • BWA(0.7.12+)
   • HTSlib(1.9)
   • Samtools(1.3.1+)
   • Bedtools(2.20.1+)
   • LUMPY
   • Python(3.5+)
   • Gurobipy
   • Python packages(3.5+)
   • pysam
   • sklearn
   • numpy
   • scipy
   • lmfit
   • ssw-py

Install HTSlib 1.9 version

See https://github.com/samtools/htslib

Install LEMON

Download and install LEMON

git clone --recursive https://github.com/lichen2018/hgt-detection.git
cd getAccBkp
make
export CPLUS_INCLUDE_PATH=/home/your_home_path/lib/htslib-1.9

LEMON usage

1. Detect raw HGT breakpoints.

usage: python LEMON/Scripts/get_raw_bkp.py [options]

Required arguments

  -r     FILE  Metagenomic Reference 
  -u     FILE  unique reads bam file
  -o     FILE  raw breakpoints file

Option arguments

  -t INT  number of threads [4]

2. Detect accurate HGT breakpoints.

usage: LEMON/getAccBkp/get_acc_bkp [options]

Required arguments

  -r        FILE  Metagenomic Reference
  -u        FILE  unique reads bam file
  -s        FILE  split reads bam file
  -b        FILE  raw breakpoints file
  -o        FILE  accurate reakpoints file
  -t        INT  number of threads 

3. Reconstruct HGT strain.

usage: python LEMON/Scripts/reconstruct_HGT_strain.py [options]

Required arguments

  -r        FILE  Metagenomic Reference
  -c        FILE  coverage file
  -s        STR   Sample name or id
  -a        FILE  accurate reakpoints file
  -o        STR   path to the directory where results should be stored

4. Reconstruct HGT strain with replication time.

usage: python LEMON/Scripts/reconstruct_HGT_strain_for_replication_time.py [options]

Required arguments

  -r        FILE  Metagenomic Reference
  -c        FILE  coverage file
  -s        STR   Sample name or id
  -a        FILE  accurate reakpoints file
  -o        STR   path to the directory where results should be stored

Example workflow

Preprocessing

# Align the data
bwa mem -M -t 8 -R "@RG\tID:id\tSM:sample\tLB:lib" Metagenomic_reference.fasta sample.1.fq sample.2.fq \
  | samtools view -bhS -> sample.unsort.bam

# Sort bam file
samtools sort -o sample.bam sample.unsort.bam

# Extract split reads
samtools view -h sample.bam \
  | lumpy-sv/scripts/extractSplitReads_BwaMem -i stdin \
  | samtools view -Sb > sample.unsort.splitters.bam

# Sort split reads bam file
samtools sort -o sample.splitters.bam sample.unsort.splitters.bam

# Extract unique reads bam file
samtools view -q 20 -b sample.bam > sample.unique.bam

# Calculate coverage
bedtools genomecov -ibam sample.bam -bg > sample.coverage.txt

Running LEMON

# 1. Detect raw HGT breakpoints.
python LEMON/Scripts/get_raw_bkp.py -r meta_ref.fasta -u sample.unique.bam -o sample.raw.txt

# 2. Detect accurate HGT breakpoints.
LEMON/getAccBkp/get_acc_bkp -r meta_ref.fasta -u sample.unique.bam -s sample.splitters.bam -t 10 -b sample.raw.txt -o sample.acc.txt

# 3. Reconstruct HGT strains for restoring replication timing profile.
python LEMON/Scripts/reconstruct_HGT_strain.py -c sample.coverage.txt -r meta_ref.fasta -s sample -a sample.acc.bkp.txt

# 4. Reconstruct HGT strains for restoring replication timing profile.
python LEMON/Scripts/reconstruct_HGT_strain_for_replication_time.py -c sample.coverage.txt -r meta_ref.fasta -s sample -a sample.acc.bkp.txt