NanoFungiBarcode

Software requirements

Dorado basecaller
- see installation
snake_make_ont_consensus_build
- will require the conda package manager
chopper quality filtering
- can be installed via conda
  - conda create -c bioconda -n chopper chopper
  - conda activate chopper

General procedure

Convert raw signal from nanopore device into a sequence of nucleotides. This step is called basecalling and is very computationally demanding.
If multiple samples have been sequenced together at the same time (multiplexing), the data will need to be demultiplexed to separate sequences from each sample by identifying a barcode adapter.
Low quality sequences should be removed before building consensus sequences. We can select only the sequences that meet a minimum mean quality score threshold (in this case Q15 on the Phred scale).
Recover consensus sequences for each sample.

0 - Prerequisites

Raw Oxford Nanopore data (in pod5 file format)
- if your data is in fast5 format, convert to pod5 here: https://pod5.nanoporetech.com/
- Record the directory of your raw data, I will refer to this directory as "raw-data/pod5-dir".
The barcoding/adapter kit used
- In this case we are using the SQK-RBK114-24 (Rapid Barcoding Kit 24 V14).
- Run dorado demux --help to see the list of kits for the --kit-name option.
An output directory to store the results. I will refer to this directory as "output-dir".
```
mkdir "output-dir"
```

1 - Basecalling

Perform the basecalling on your raw pod5 files in the raw-data/pod5-dir directory.

dorado basecaller \
    --no-trim \
    --recursive \
    --device cpu \
    sup "raw-data/pod5-dir" \
    > output-dir/basecalled_reads.bam

The --no-trim option tells dorado to not perform any adapter trimming (which will be done in the demultiplexing step).

The --recursive option tells dorado to look through all the subfolders of the raw-data/pod5-dir for pod5 files.

The --device cpu option tells dorado to use the CPU device to perform the basecalling. If you have a CUDA enabled GPU or Apple M1/M2 installed run dorado basecaller --help for other options. CPU basecalling will be much slower than GPU basecalling.

The sup option tells dorado to use the 'super accurate' basecalling model which will produce the highest accuracy sequences at high computational cost. See which models are available here.

2 - Demultiplexing

Make a directory to store the demultiplexed sequences.

mkdir "ouput-dir/demux"

Ensure the barcoding kit name (--kit-name) is the same as the one used in the library preparation or else most sequences will be unclassified.

dorado demux \
  --kit-name SQK-RBK114-24 \
  --output-dir output-dir/demux \
  --emit-fastq \
  output-dir/basecalled_reads.bam

The --emit-fastq option tells dorado to output .fastq formatted files. It will output .bam files by default (without this option).

At the end of this step there should be one file per sample in the output-dir/demux directory. Sequences where the barcode could not be classified will be placed in output-dir/demux/unclassified.fastq.

3 - Quality filtering

This command iterates through the .fastq file for each sample and removes sequences that have a mean quality score lower than Q15.

for filename in output-dir/demux/*.fastq; do \
  cat ${filename} | chopper --quality 15 \
    > output-dir/filt-Q15/$(basename $filename);\
done

4 - Building a polished consensus sequence

snake_make_ont_consensus_build is a snakemake pipeline that can do this. Details on how this pipeline works can be found here: https://github.com/ritatam/nanopore-biosecurity-workshop-2022

Follow the steps on how to run the pipeline here: https://github.com/ritatam/snakemake_ont_consensus_build . They have been repeated here for completeness:

Clone the repository

 git clone https://github.com/ritatam/snakemake_ont_consensus_build.git
 cd snakemake_ont_consensus_build

Configure paths in the config.yaml file. These paths are relative to the config.yaml file.

input_reads_dir: "../output-dir/filt-Q15"
output_dir: "../output-dir/ont_consensus"

Install snakemake into a conda environment

conda install snakemake

Install dependencies into a conda environment

snakemake --use-conda --conda-create-envs-only -c1

Do a dry-run of the pipeline:

snakemake --use-conda -np

Run the pipeline:

snakemake --use-conda -c8

Royal-Botanic-Gardens-Victoria/NanoFungiBarcode