ABO blood typing using Oxford Nanopore MinION sequencing

ABO sequences were aquired from the NCBI dbRBC database:

https://www.ncbi.nlm.nih.gov/projects/gv/mhc/xslcgi.cgi?cmd=bgmut/home

See https://ftp.ncbi.nlm.nih.gov/pub/mhc/rbc/Final%20Archive/Excel_and_PowerPoint/ for some literature.

Required tools

The pipeline makes use of the following core dependencies:

- bioconda::fastqc=0.12.1
- bioconda::bwa=0.7.17
- conda-forge::ncurses
- bioconda::samtools=1.19.2
- bioconda::minimap2=2.26
- conda-forge::biopython=1.83
- python=3.10
- pip
- pip:
  - numpy>=1.26.0
  - Bio>=1.6.0
  - biopython>=1.8o
  - openpyxl>=3.1.0
  - pandas>=2.2.0
  - pysam>=0.22.0
  - matplotlib>=3.8.0
  - XlsxWriter>=3.2.0
  - multiqc>=1.18

A complete list of dependencies is found in the assets folder assets/conda.yml.

Required input files structure

Ensure that all input fastq files have a naming convention that matches this regular expression (regex)

## python regex for matching samples
pattern = r"^(IMM|INGS|NGS|[A-Z0-9]+)(-[0-9]+-[0-9]+)?_barcode\d+$"

The regex does the following:

• ^(IMM|INGS|NGS|[A-Z0-9]+) allows for files strating with the prefixes IMM, INGS, NGS, or any combination of letters A-to-Z and digits 0-to-9.
• (-[0-9]+-[0-9]+)? handles optional segments of digits separated by a dash(-).
• _barcode\d+$ ensures the filename ends with_barcode followed by digits to denote barcode numbers.

There is a filename handling logic in the code filename.split("_") assumes that the barcode is always the last part of the filename. The names are split into basename and barcode which are then used in later reporting. Please Adjust this if necessary based on actual filename structure in your assays.

Here are a few examples of acceptable input file names:

NGSPOS_barcode13.fastq
NGSNEG_barcode12.fastq
INGSPOS_barcode01.fastq
INGSNEG_barcode96.fastq
IBTGSPOS_barcode19.fastq
2025705_barcode14.fastq
IMM-24-44988_barcode51.fastq
Sample1-2024-12345_barcode22.fastq

Testing without nextflow

The pipeline can be tested of single input file by cloning this repo and installing all dependncies above, then running the following commands:

python bin/AnalyzeAbo_Main.py  \
  --reference="assets/A1_01_01_1_reference_Exon6.fasta" \
  --alleles="assets/ABO_Database.fasta" \
  --output="SampleName/exon6" \
  --analysis-type="READS" \
  --reads="SampleName.fastq" \
 
 python bin/AnalyzeAbo_Main.py  \
  --reference="assets/reads_bc51/A1_01_01_1_reference_Exon7.fasta" \
  --alleles="assets/input/ABO_Database.fasta" \
  --output="SampleName/exon7" \
  --analysis-type="READS" \
  --reads="SampleName.fastq"

Looping through a couple of samples with the above command will generate the following outputs:

Output data structure from the testing

OutputDirectoryName/
├── Sample1
│   ├── exon6
│   │   └── alignment
│   └── exon7
│       └── alignment
├── Sample2
│   ├── exon6
│   │   └── alignment
│   └── exon7
│       └── alignment
├── Sample3
│   ├── exon6
│   │   └── alignment
│   └── exon7
│       └── alignment

With individual files named as follows:

OutputDirectoryName/
├── Sample1
│   ├── exon6
│   │   ├── ABOPhenotype.txt
│   │   ├── ABOReadPolymorphisms.txt
│   │   ├── alignment
│   │   │   ├── alignment.bam
│   │   │   ├── alignment.bam.bai
│   │   │   ├── AlignmentReference.fasta
│   │   │   ├── AlignmentReference.fasta.amb
│   │   │   ├── AlignmentReference.fasta.ann
│   │   │   ├── AlignmentReference.fasta.bwt
│   │   │   ├── AlignmentReference.fasta.pac
│   │   │   └── AlignmentReference.fasta.sa
│   │   └── ReadAlignmentSpreadsheet.csv
│   ├── exon7
│   │   ├── ABOPhenotype.txt
│   │   ├── ABOReadPolymorphisms.txt
│   │   ├── alignment
│   │   │   ├── alignment.bam
│   │   │   ├── alignment.bam.bai
│   │   │   ├── AlignmentReference.fasta
│   │   │   ├── AlignmentReference.fasta.amb
│   │   │   ├── AlignmentReference.fasta.ann
│   │   │   ├── AlignmentReference.fasta.bwt
│   │   │   ├── AlignmentReference.fasta.pac
│   │   │   └── AlignmentReference.fasta.sa
│   │   └── ReadAlignmentSpreadsheet.csv
│   ├── Sample1_exon6.log.txt
│   └── Sample1_exon7.log.txt

The ABOPhenotype.txt files from each sampe can then be collated using:

python bin/Aggregate_ABO_reports.py OutputDirectoryName

The nextflow workflow

The steps above are simplified in a NextFlow; https://www.nextflow.io/ pipeline that does all the above steps and streamlines installation of requisite software and tools with a single command.

Besides reproducability, nextflow offeres several advatages over conventional for loops, including scallability, portability, and debugging/resumption of failed tasks.

Input files and output directory can be defined in the config files or provided directly in the commandline.

To analyse files with config, run:

• nextflow run main.nf -resume (user can override inputs and output using --reads '*.fastq' --outdir 'ABO_results' on the commandline).

We have also added the ability for the pipeline to automatically set-up a conda or docker based environment with all required tools and libraries.

Users may also opt for a workload manager such as -profile slurm,docker|-profile slurm,conda, is which case, all required modules docker/conda must be installed and loaded. The config slurm parameters must also be defined to ensure tasks are submitted to the correct resource queue/account.

For conda environment, it is advisable to prepare the working computer using mamba for easy resolution of environments. Follow these steps to achieve better results.

mamba create -y -n abo-analysis-env
conda activate abo-analysis-env
mamba env update --file abo-analysis/assets/conda.yml --prune
conda deactivate

# If conda cativate fails, run:
source {path_to_anaconda}/anaconda3/etc/profile.d/conda.sh

To run without the workload manager but with a specific containerization, use:

• nextflow run abo-analysis/main.nf -resume --outdir "$PWD/230128R_ABO_results" -with-conda abo-analysis-env or nextflow run abo-analysis/main.nf -resume --outdir "$PWD/230128R_ABO_results" -profile conda
• nextflow run abo-analysis/main.nf -resume --outdir "$PWD/230128R_ABO_results" -with-docker fmobegi/abo-analysis or nextflow run abo-analysis/main.nf -resume --outdir "$PWD/230128R_ABO_results" -profile docker

Renaming samples

The code by default renames samples using a tab file with sequencingID and sampleName (see nextflow.config file under $params.renaming_file). This option is controlled by the parameter $params.skip_renaming and can be overridden via the commandline using option --skip_renaming true to skip the process.

Results from the Nextflow pipeline will look something like this

230128R_ABO_results/
├── ABO_result.txt
├── ABO_result.xlsx
├── execution_report.html
├── execution_timeline.html
├── execution_trace.txt
├── SampleName
│   ├── exon6
│   │   ├── ABOPhenotype.txt
│   │   ├── ABOReadPolymorphisms.txt
│   │   ├── alignment
│   │   │   ├── alignment.bam
│   │   │   ├── alignment.bam.bai
│   │   │   ├── AlignmentReference.fasta
│   │   │   ├── AlignmentReference.fasta.amb
│   │   │   ├── AlignmentReference.fasta.ann
│   │   │   ├── AlignmentReference.fasta.bwt
│   │   │   ├── AlignmentReference.fasta.pac
│   │   │   └── AlignmentReference.fasta.sa
│   │   └── ReadAlignmentSpreadsheet.csv
│   ├── exon7
│   │   ├── ABOPhenotype.txt
│   │   ├── ABOReadPolymorphisms.txt
│   │   ├── alignment
│   │   │   ├── alignment.bam
│   │   │   ├── alignment.bam.bai
│   │   │   ├── AlignmentReference.fasta
│   │   │   ├── AlignmentReference.fasta.amb
│   │   │   ├── AlignmentReference.fasta.ann
│   │   │   ├── AlignmentReference.fasta.bwt
│   │   │   ├── AlignmentReference.fasta.pac
│   │   │   └── AlignmentReference.fasta.sa
│   │   └── ReadAlignmentSpreadsheet.csv
│   ├── SampleName_exon6.log.txt
│   └── SampleName_exon7.log.txt
├── software_versions.txt
└── workflow.oncomplete.txt

Feel free to raise an issue or reach out if you need any support getting this tool running, or with suggestions for improvement.