RecallME is a containerized python tool for VCF file comparison and variant calling pipelines benchmarking and optimization.
To run RecallME you need to activate the conda environment and to install singularity (> v.3) and ANNOVAR (it requires a license free of charge for non-commercial purposes). For more information about ANNOVAR, please visit the official website
To start using RecallME v.0.2, type the following commands:
git clone https://github.com/mazzalab-ieo/recallme.git
cd recallme/
conda env create --file env.txt --name {name_env}
To run a comparison, type:
python RecallME.py -q $QUERY_VCF \
--high_conf_bed $BED \
--query_format VCF \
-g $GROUND_TRUTH \
--gt_format VCF \
-f $FASTA \
--genome_bed $GENOME_BED \
-b $BAM \
-a $ANNOVAR_DIR \
-o $OUT_DIR \
--caller $CALLER_NAME
- GENOME_BED is a bedtools specific bed file with chromosome sizes in bps (i.e. chr1 n) or a .fai index.
- BAM file for query
- ANNOVAR_DIR the directory where is stored convert2annovar.pl which is required for running RecallME
- CALLER_NAME the name of the caller used to produce the query VCF file (GATK, TVC, Deepvariant, VarScan, LoFreq)
For more info on required and not required commands, please type:
python RecallME.py --help
To run the example, open a terminal and type:
conda activate {name_env}
cd recallme/
#add the annovar folder within run_example.sh
bash run_example.sh
run_example.sh produces an outdir/ folder in the run_example/ directory as follows:
After the output has been generated, go to the web app and load the AVinputs (query_split_var_type.avinput + ground_truth_var_type.avinput) and the rds object to visualize your results and to optimize your pipeline.
For additional information, please check the about page for interrogating the manual.
RecallME is under MIT License. BEDTOOLS is under GNU General Public License.
If you use RecallME in a scientific article, please cite: G Vozza, E Bonetti, G Tini, V Favalli, G Frige’, G Bucci, S De Summa, M Zanfardino, F Zapelloni, L Mazzarella, Benchmarking and improving the performance of variant-calling pipelines with RecallME, Bioinformatics, 2023; doi: https://doi.org/10.1093/bioinformatics/btad722