Assembly polishing is done to correct any errors left in the assembly by using DNA short reads as a reference. It is done in two stages. First, the trimmed DNA short reads are mapped onto the genome assemblies produced by Flye, and then the assemblies are polished from the results of mapping.

• BWA - BWA is a fast and memory-efficient tool for aligning short sequencing reads to a reference genome.
• samtools - samtools is a suite of tools, used to work with SAM/BAM files which are prominently seen with read alignments.

• .fasta files of the assemblies generated from Flye and Canu.
• .fastq.gz files of the processed DNA short reads.

First up, the assemblies have to be prepared by indexing the assembly into a compressed, searchable version of the assembly. This makes it easier for BWA to find where each short read might align in the assembly, down the line.

#load the necessary modules

module load bioinfo-tools
module load bwa
module load samtools

#do indexing 

bwa index assembly.fasta

This step produces 5 different files - .amb, ann, .bwt, .pac, .sa. All of these files will be used by BWA during alignment. Now, the reads can be aligned to the assembly. Here BWA takes each read pair and tries to find the best match on the assembly, as it compares the reads to the index and calculates where the read likely originated. The results are saved in a .sam file.

bwa mem assembly.fasta read_1.fastq.gz read_2.fastq.gz > aligned_reads.sam

Then the resulting .sam file is converted to a .bam file, as it makes the file smaller and easier to process. The .bam file is also sorted and indexed into a .bam.bai file, as the .bam.bai index file make it easier for the downstream tools to access regions of .bam quickly, rather than parsing through the file every time.

samtools view -S -b aligned_reads.sam > aligned_reads.bam
samtools sort aligned_reads.bam -o aligned_reads_sorted.bam
samtools index aligned_reads_sorted.bam

• Pilon - Pilon uses aligned short reads to fix base errors, correct indels, fill gaps, improve sequence accuracy etc., It works by checking how well the reads match the assembly, then makes corrections based on consensus.

• .fasta files of the assemblies generated from Flye and Canu.
• .bam and .bam.bai files generated from mapping.

Pilon goes through the assembly and then the aligned reads in .bam file to scan for mismatches, low-confidence regions etc. If a lot of reads disagree with a specific position in the assembly, Pilon corrects it.

module load bioinfo-tools
module load pilon

java -Xmx16G -jar $PILON_HOME/pilon.jar \
  --genome assembly.fasta \
  --frags aligned_reads_sorted.bam \
  --output polished_assembly \
  --threads 8

.fasta file of the polished assemblies.