Pectobacterium and microbiome analysis of potato stem samples

Please download or obtain access to the following before doing the analysis

Programs

Clean metagenomic raw reads

• Trimmomatic

Metagenomic analysis

• Kraken2
• Sourmash
• Pectobacterium custom-made Database

Map host genome and remove mapped reads

• BWA
• Samtools

Illumina QC

Trimmomatic

Trimmomatic cleans the metagenomic Illumina reads based on quality control. Trimmomatic uses Java, therefore for our script, we use the module system to use it.

module load jdk/1.8.0
for i in $(ls *.fastq | grep "_R1" | cut -f 1 -d "_"); do \
N=$(basename $i .fastq); java -jar trimmomatic-0.39.jar PE -phred33 \
#Seperate files into R1 and R2
${i}_R1.fastq ${i}_R2.fastq \
/cleanreads/$N.R1.paired.fastq
/cleanreads/$N.R1.unpaired.fastq \
/cleanreads/$N.R2.paired.fastq
/cleanreads/$N.R2.unpaired.fastq \
ILLUMINACLIP:NexteraPE-PE.fa:2:30:10 \
LEADING:3 TRAILING:3 SLIDINGWINDOW:4:15 MINLEN:100;done

#We use /cleanreads/ as the outputfolder
#make sure your PATH for your raw data does not contain underscores as that will be cut and the script will not work.

Metagenome Analysis

Kraken - Create a database using NCBI genomes

1. Download Taxonomy to create a database called DB

cd /kraken2/
./kraken2-build --download-taxonomy --db /folder/DB

2. Add genomes to the database library

#keep all NCBI genomes in the same folder (genomes/) to create a library with a loop.

for files in /genomes/*.fna; do \
./kraken2-build --add-to-library $file --db /folder/DB; done

3. Build the database

./kraken2-build --build --db /folder/DB

Kraken - run the analysis

#input-location/ is a folder that contains all the clean metagenomic raw reads.

cd /kraken2
for i in $(ls /input-location/*.fastq | grep "_R1" | cut -f 1 -d "_"); do \
N=$(basename $i .fastq); ./kraken2  \
--db /folder/DB \
--paired --report $N.report \
--output $N.kraken \
${i}_R1.fastq ${i}_R2.fastq; done

Sourmash

Create a collection of signatures (a database)

1. Sourmash sketch to create the signature

#genomes is a folder that 
sourmash sketch dna -p k=31 genomes/*.fna -outdir genomesDB

2. Index the DB

sourmash index -k 31 genomesDB.sbt.zip genomesDB/*.sig

Create signatures for the metagenomic reads

for i in $(ls metagenomic_reads/*.fq | grep "_R1" | cut -f 1 -d "_"); do \
N=$(basename $i .fq);sourmash sketch dna -p k=31 ${i}_R1.fq ${i}_R2.fq --name $N \
-o $N.zip;done

Compare DB signatures with the metagenomics reads

for i in $ $N.zip; do N=$(basename $i .zip); sourmash gather -p k=31 $i genomesDB.sbt.zip

Remove host reads before metagenomic analysis

Another option is to remove the host (potato) reads before the analysis

BWA-mem

module load bwa/0.7.17
bwa index hostgenome.fa \

for i in $(ls metagenomicreads/*.fastq | grep "_R1" | cut -f 1 -d "_"); do \
N=$(basename $i .fastq); bwa mem hostgenome.fa ${i}_R1.fastq ${i}_R2.fastq \
> $N.sam; done

Samtools

The unmapped reads to the potato genome can be extracted from the SAM file. The unmapped reads correspond to the microbial reads.

1. convert to bam files and sort

cd /output-location
for i in /bwamem-output-location/*.sam; do \
N=$(basename $i .sam); samtools sort $i -O bam -o $N.bam; done

2. Keeping unmapped reads

for i in *.bam; do \
N=$(basename $i .bam); samtools view -b -f 4 $i -o $N.unmapped.bam; done

3. sort unmapped

for i in *.unmapped.bam; do \
N=$(basename $i .unmapped.bam); samtools sort $i -n -o $N.sorted.bam; done

4. convert bam to fastq files

for i in *.sorted.bam; do \
N=$(basename $i .sorted.bam); samtools fastq $i > $N.nonpotato.fastq; done