SWAAT allows for the prediction of the variant effect based on structural properties. SWAAT annotates a panel of 36 ADME genes including 22 out of the 23 clinically important members identified by the PharmVar consortium. The workflow consists of a set of python codes. The execution of these code is managed within Nextflow to annotate coding variants based on 37 criteria. SWAAT also includes an auxiliary workflow allowing a versatile use for genes other than the ADME members. The auxiliary workflow can be found in the directory auxiliary_wf of the repository.

If you use SWAAT, please cite the following article

Othman H, Jemimah S, da Rocha JEB. SWAAT Bioinformatics Workflow for Protein Structure-Based Annotation of ADME Gene Variants. Journal of Personalized Medicine. 2022; 12(2):263. https://doi.org/10.3390/jpm12020263

1. Installing dependencies

1.1 Dependencies for the main workflow

• Python 3.7.4
• Nextflow 20.10.0
• Biopython 1.76
• numpy 1.18.1
• pandas 1.0.4
• scikit-learn 0.22.2
• scipy 1.4.1
• FoldX
• ENCoM
• freesasa 2.0.3
• Stride

1.2 Instruction for installing the dependencies

All the dependencies are open source. The user needs only to obtain a free licence for FoldX. We recommend installing the dependencies for SWAAT using conda virtual environment whenever it's possible. First let's create a virtual environment called SWAAT and install all the python dependencies including biopython, numpy, pandas, scikit-learn and scipy. We provided a YAML file in the root directory to automate the process. We also recommend using the Mamba package manager which will speed up the building process.

First, install Anaconda or miniconda to be able to create and manage virtual environments.

To create and build an environment called SWAAT use the following command:

conda env create -f SWAAT_dependencies.yml

Once the environment is built you can activate it using the following command:

conda activate SWAAT

1.3 Special consideration for Installing ENCoM

ENCoM can be obtained from its git repository. The build_encom code systematically returns an exit status of 1 which creates an issue when running within a Nextflow process. To overcome the problem, you need to modify line 370 from return(1); to return(0); in src/build_encom.c. Thereafter you can compile the set of codes as following:

cd ENCoM/build/ \
	&& make -f Makefile.Unix

2. Running SWAAT

First, you need to clone the SWAAT repository from GitHub:

git clone https://github.com/hothman/SWAAT.git
cd SWAAT

The main workflow, that runs the annotation process is implemented in main.nf file. To list the options and the arguments type the following command in the source directory after :

nextflow main.nf --help

Arguments:
  --dbhome [folder]               Path to database containing the dependency files for annotating the variants (Default False)
  --vcfhome [folder]              Path to folder containing VCF files split by annotated gene (e.g. CYP2D6.vcf) (Default False)
  --outfolder [str]               Where to output the plain text and the HTML report (Default: false)
  --genelist [file]               User can limit the annotation to the list of genes contained in a this text file (one line per gene) (Default False)

Other
  --foldxexe [str]                Specifies the name of the executable of FoldX software (Default foldx)
  --encomexe [str]                Specifies the name of the executable of build_encom (Default build_encom)
  --freesasaexe [str]             Specifies the name of the executable of freesasa software (Default freesasa)
  --strideexe [str]               Specifies the name of the executable of stride software (Default freesasa)
  --rotabase [abs path]           Path to rotabase.txt (Default PATH to foldx)

The VCF files to annotate should be specific for each ADME gene in an uncompressed format. The file name for the VCF file must contain the gene symbol in upper case letters (e.g CYP1A2.vcf and TPMT.vcf). A full list of the annotated ADME genes is contained in ./database/gene_list/all_adme_genes.txt. The path to folder that contains the VCF files is specified in the --vcfhome argument.

A typical run of the workflow could be the following:

nextflow run main.nf --dbhome /home/hothman/SWAAT/database/ \
		--vcfhome /home/hothman/SWAAT/vcfs \
		--outfolder ./swaat_out \
		--genelist ./inputexample/gene_list.txt

A maintained database can be downloaded with the repository in ./database folder. The argument --dbhome allows to specify the path to the database.

3. SWAAT Output

SWAAT generates a CVS file, a HTML file, and a set of directories referring to each annotated gene. The output are complementary.

3.1 CSV output

The CSV file offers richer content than the HTML file and serves for elaborate analysis and filtering methods. Each column provides a piece of information according to structure and sequence-based features.

Column	information
gene_name	Gene symbol name
wt_res	Residue type in the amino acid reference sequence
mut_res	Amino acid type encoded by the genetic variant in the provided VCF file
position	Position in the amino acid reference sequence
chain	Chain identifier in the PDB reference structure
ddG	Variation of the folding energy upon mutation calculated by FoldX (kcal/mol)
SecStruc	Secondary structure type in the reference structure
ddS	Variation of the vibrational entropy upon mutation calculated from ENCom output (kcal/mol)
subScore	BLOSUM62 substitution score
grantham	Grantham score
sneath	Sneath's index
classWT and classMut	Amino acid class of the reference and the variant respectively according to Albatineh and Razeghifard (2008). c1: M, V, I, L; c2 :G, A, C, P, S, T; c3 : F, W, Y; c4: R, H, K; c5: N, Q, D, E
sasa_mut and sasa_wt	Solvent-accessible surface area of the reference and the variant structures respectively
hb_mut and hb_wt	Number of hydrogen bonds established by the reference amino acid and the variant amino acid respectively
sb_mut and sb_wt	Number of salt bridges established by the reference amino acid and the variant amino acid respectively
sasa_ratio	Ratio of accessible surface area beween the reference and the variant amino acid.
hyrophob_WT and hyrophob_Mut	Hydrophobicity of the reference and the variant amino acids respectively according to JOND750101 descriptor from AAindex database
volume_WT and volume_Mut	Amino acid volume of the reference and variant residues respectively according to BIGC670101 descriptor from AAindex database
pssm_wt and pssm_mut	PSSM scores of the reference and variant amino acids respectively.
disulfide_breakage buried_Pro_introduced buried_glycine_replaced buried_hydrophilic_introduced buried_charge_introduced buried_charge_switch sec_struct_change buried_charge_replaced buried_exposed_switch exposed_hydrophilic_introduced Buried_salt_bridge_breakage Large_helical_penality_in_alpha_helix	Indicates if a significant structural change occured upon the mutation according to Missense3D criteria, 1: True, 0: False
swaat_prediction	Predictive model outcome, 1: significant impact, 0: Neutral

3.2 HTML output

The HTML report is better in summarizing significant structural events with likely significant impact. These can be spotted as red flag tags. The report also links the genome coordinates with the amino acid substitution. In addition, the report offers the sequence annotation associated with each amino acid position as well as some statistics about the drug-binding likelihood of the amino acid. These statistics were calculated from the FTMap output which reports the number of probe molecules that likely bind to the amino acid.

4. Example: annotattion of six ADME genes

The example shows the annotation process of seven ADME genes CYP2A13 , CYP2B6 , CYP2C19. CYP2D6. CYP3A4, and CYP1A2. We have optained the VCF files of these gene from PharmVar and store them in inputexample directory. We generated a list of all the gene symbols corresponding to the VCF files in ./inputexample/gene_list.txt shown below:

CYP2A13
CYP2B6
CYP2C19
CYP2D6
CYP3A4
CYP1A2

The gene list will tell SWAAT to limit the annotation only for the gene listed in the file.

In this case the user must specify the location of the VCF files, the location of the gene list file with an option to specify the path and name of the output folder. Here the final result is saved in swaat_output. The workflow can then be run from the command line.

nextflow run main.nf --dbhome /home/hothman//SWAAT/database/ \
		--vcfhome /home/hothman/SWAAT/inputexample/VCFs \
		--outfolder ./swaat_out \
		--genelist ./inputexample/gene_list.txt

The workflow will process 113 variants in total.

---

Funding

This work was primarily funded through a grant by GlaxoSmithKline Research and Development Ltd to the Wits Health Consortium. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

Acknowledgement

We thank Scott Hazelhurst and Nhlamulo Khoza from the University of the Witwatersrand who both spent significant effort to improve and test the tools developed in this project.

Authors

Houcemeddine Othman, Sherlyn Jemimah and Jorge EB da Rocha.

Please contact Houcemeddine Othman ([email protected]/[email protected]) for any questions or comments.