Extracting unmapped sequences

Here we describe the process to fetch and extract unmapped reads onto a reference genome(s)/transcriptome using BWA.
This guide assumes that conda is already installed - if you have not yet done so, see here.

Pre-requisites

Installed conda as per Installing on Linux — conda 24.11.4.dev30 documentation
Raw FASTQ files have been quality trimmed (i.e., remove adapters and poor quality passes (i.e., >=Q30)).

Strategy #1: BWA aligner

STEP 1.1 - (Optional) Create a conda environment with necessary tools if not yet available

Creating conda environment, and you can give it any name (i.e., related to the workflow or tool being installed). For simplicity, here we will call it ‘myanalyses’

conda create --name myanalyses python=3.7

Activate the conda environment.

conda activate myanalyses

Let’s now necessary tools one at a time (faster than installing them all at once):

conda install -c bioconda bwa

To look for other bioinformatics tools to install, search the tool of interest at: https://anaconda.org

STEP 1.2 - Mapping of reads onto reference sequences (i.e., can be abundant genomic sequences found in an environmental sample).

Here we aim to generate a mapped file (BAM) that can be used to identify unmapped sequences of interest to run a follow-up de novo genome assembly, for example.
See BWA user manual for more details: https://bio-bwa.sourceforge.net/bwa.shtml
For the mapping, we can use either individual FASTQ files for each sample or merged files for a group of related samples - if you merge the FASTQ files, these need to be merged using the sample order for both READ1 (*R1.fastq) and READ2 (*R2.fastq) files. We assume that adaptors have already been removed from the raw reads and the individual or merged files are quality processed reads (for example, processed by either cutadapt, trimmomatic, trim galore, fastp or another QC tool)

Strategy #2: bowtie aligner

STEP 2.1 - (Optional) Create a conda environment with necessary tools if not yet available

(optional) If not yet created as per above - Create a conda environment, and you can give it any name (i.e., related to the workflow or tool being installed). For simplicity, here we will call it ‘myanalyses’

Activate the conda environment.

Let’s now necessary tools one at a time (faster than installing them all at once):

STEP 1.2 - Mapping of reads onto reference sequences (i.e., can be abundant genomic sequences found in an environmental sample).

Here we aim to generate a mapped file (BAM) that can be used to identify unmapped sequences of interest to run a follow-up de novo genome assembly, for example.
See bowtie user manual for more details: https://bowtie-bio.sourceforge.net/manual.shtml
For the mapping, we can use either individual FASTQ files for each sample or merged files for a group of related samples - if you merge the FASTQ files, these need to be merged using the sample order for both READ1 (*R1.fastq) and READ2 (*R2.fastq) files. We assume that adaptors have already been removed from the raw reads and the individual or merged files are quality processed reads (for example, processed by either cutadapt, trimmomatic, trim galore, fastp, or another QC tool)

Build bowtie-index for ref genomes:

Run mapping allowing no mismatches (identity = 100%)