Automating ACMG Variant Classifications With BIAS-2015 v2.0.0: Algorithm Analysis and Benchmark Against the FDA-Approved eRepo Dataset

Background In 2015, the American College of Medical Genetics and Genomics (ACMG), in collaboration with the Association of Molecular Pathologists (AMP), published guidelines for interpreting and classifying germline genomic variants. These guidelines defined five categories: benign, likely benign, uncertain significance, likely pathogenic, and pathogenic, with 28 criteria but no specific implementation algorithms.

Here we present Bitscopic Interpreting ACMG Standards 2015 (BIAS-2015 v2.0.0), an open-source software that automates the classification of variants based on 19 ACMG criteria while enabling user-defined weighting and manual adjustments for clinical contexts. BIAS-2015 supports high-throughput classification via command line, along with a web-based GUI, enabling variant review, modification, and interactive curation. Using genomic data from the FDA-recognized ClinGen Evidence Repository (eRepo v2.2.0), we evaluated BIAS-2015’s sensitivity, specificity, and F1 values with expert curation. BIAS-2015 demonstrated superior performance to InterVar, achieving a pathogenic sensitivity of 73.99% (vs. 64.31%), benign sensitivity of 80.23% (vs. 53.91%), and a 11x speed improvement, classifying 1,327 variants per second.

BIAS-2015 provides an accurate, scalable, and transparent ACMG classification framework. All code and the interactive variant curation platform are available on GitHub. Competing Interest Statement The authors declare the following conflict of interest: The BIAS-2015 algorithm described in this paper is intended for use in a commercial product by Bitscopic. This potential financial interest does not alter the authors adherence to community best practices and scientific standards in developing and validating the algorithm. The development of BIAS-2015 was conducted independently by Bitscopic, and no external funding or influence was received from commercial entities outside of this potential interest. Funding Statement The authors declare the following conflict of interest: The BIAS-2015 algorithm described in this paper is intended for use in a commercial product by Bitscopic. This potential financial interest does not alter the authors adherence to community best practices and scientific standards in developing and validating the algorithm. The development of BIAS-2015 was conducted independently by Bitscopic, and no external funding or influence was received from commercial entities outside of this potential interest. Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes The details of the IRB/oversight body that provided approval or exemption for the research described are given below: The study used ONLY openly available human data that were originally located at: UCSC Genome Browser public datasets: RefSeq: https://hgdownload.soe.ucsc.edu/goldenPath/hg19/database/ncbiRefSeqHgmd.txt.gz RepeatMasker: https://hgdownload.soe.ucsc.edu/goldenPath/hg19/database/rmsk.txt.gz CCDS: https://hgdownload.soe.ucsc.edu/goldenPath/hg19/database/ccdsGene.txt.gz GWAS Catalog: https://hgdownload.soe.ucsc.edu/goldenPath/hg19/database/gwasCatalog.txt.gz AbSplice: https://hgdownload.soe.ucsc.edu/gbdb/hg19/abSplice/AbSplice.bb AVADA: http://hgdownload.soe.ucsc.edu/gbdb/hg19/bbi/avada.bb UniProt Domains: https://hgdownload.soe.ucsc.edu/gbdb/hg19/uniprot/unipDomain.bb gnomAD Missense Constraint: https://hgdownload.soe.ucsc.edu/gbdb/hg19/gnomAD/missense/missenseConstrained.bb ClinVar dataset (NCBI): https://ftp.ncbi.nlm.nih.gov/pub/clinvar/vcf_GRCh37/clinvar.vcf.gz Illumina Nirvana (v3.18.1): https://github.com/Illumina/Nirvana/releases/tag/v3.18.1 FDA-approved eRepo database (ClinGen Evidence Repository): https://erepo.clinicalgenome.org/evrepo/ All datasets used in this study are fully public, do not require any registration, licensing, or access approval, and were openly available before the initiation of the study. I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable. Yes Data Availability The algorithm code, pre-processing scripts, validation tools, image generation code, and manuscript creation resources are openly available on GitHub to ensure transparency and accessibility for the broader scientific community. Algorithm Code and Resources: https://github.com/bitscopic/BIAS-2015 GUI Code: https://github.com/bitscopic/BIAS-2015-UI Precomputed required files for BIAS-2015 v2.0.0 are available via AWS S3: HG19 Data Files: s3://bias-2015/bias_v2.0.0_hg19_data_files.zip HG38 Data Files: s3://bias-2015/bias_v2.0.0_hg38_data_files.zip All supplementary files, including the validation VCF and intermediate validation files used in the manuscript, are available at: s3://bias-2015/bias_v2.0.0_supplementary_files.zip Data Availability Statement: All data used in this study are openly available through the sources listed above. No restrictions apply to data access.