Open-Sourced CIViC Annotation Pipeline to Identify and Annotate Clinically Relevant Variants Using Single-Molecule Molecular Inversion Probes

Abstract

Purpose: Clinical targeted sequencing panels are important for identifying actionable variants for patients with cancer; however, existing approaches do not provide transparent and rationally designed clinical panels to accommodate the rapidly growing knowledge within oncology.

Materials and methods: We used the Clinical Interpretations of Variants in Cancer (CIViC) database to develop an Open-Sourced CIViC Annotation Pipeline (OpenCAP). OpenCAP provides methods to identify variants within the CIViC database, build probes for variant capture, use probes on prospective samples, and link somatic variants to CIViC clinical relevance statements. OpenCAP was tested using a single-molecule molecular inversion probe (smMIP) capture design on 27 cancer samples from 5 tumor types. In total, 2,027 smMIPs were designed to target 111 eligible CIViC variants (61.5 kb of genomic space).

Results: When compared with orthogonal sequencing, CIViC smMIP sequencing demonstrated a 95% sensitivity for variant detection (n = 61 of 64 variants). Variant allele frequencies for variants identified on both sequencing platforms were highly concordant (Pearson's r = 0.885; n = 61 variants). Moreover, for individuals with paired tumor and normal samples (n = 12), 182 clinically relevant variants missed by orthogonal sequencing were discovered by CIViC smMIP sequencing.

Conclusion: The OpenCAP design paradigm demonstrates the utility of an open-source and open-access database built on attendant community contributions with peer-reviewed interpretations. Use of a public repository for variant identification, probe development, and variant interpretation provides a transparent approach to build dynamic next-generation sequencing-based oncology panels.

FIG 1.

Methods for Clinical Interpretations of Variants in Cancer (CIViC) single-molecule molecular inversion probe (smMIP) development and validation using the Open-Sourced CIViC Annotation Pipeline (OpenCAP). The first series describes CIViC smMIP development. Variants were selected using sequence ontology identification numbers (IDs) and the CIViC Variant Evidence Score. Subsequently, eligible variants were categorized based on length, and smMIP reagents were designed to target regions of interest. The second series describes sample selection and sequencing methods. In total, there were 22 tumor samples derived from 5 tumor subtypes. Of these 27 samples, 15 had tumor and paired normal samples, and 7 were tumor-only samples. The third series shows the analysis used to validate the CIViC smMIP design. Variants were called using the pipeline described in Materials and Methods, and accuracy was attained by comparing variants observed on original sequencing to variants observed using the CIViC smMIP capture panel. Variant allele frequencies (VAFs) across both platforms were also compared. AML, acute myeloid leukemia; CRC, colorectal cancer; HL, Hodgkin lymphoma; HNSCC, head and neck squamous cell carcinoma; SCLC, small-cell lung cancer.

FIG 2.

Regions targeted by the Clinical Interpretations of Variants in Cancer (CIViC) single-molecule molecular inversion probes (smMIPs) are, by design, supported by extensive clinical relevance according to the CIViC database. Variants that were eligible for CIViC smMIP development were divided into various coverage methods based on sequence ontology identification number and length. The bar graph shows the total number of evidence items used for each of the groups parsed by the evidence type.

FIG 3.

Waterfall plot showing extensive overlap between variants observed using original exome or whole-genome sequencing with variants observed using Clinical Interpretations of Variants in Cancer (CIViC) single-molecule molecular inversion probe (smMIP) sequencing. Each column represents a sample that had original exome or whole-genome sequencing with subsequent orthogonal validation using the CIViC smMIP sequencing. Rows represent mutated genes across all samples. Numbers within each box represent the variant allele frequency (VAF) observed on original exome or whole-genome sequencing. Red boxes indicate that a variant was observed by CIViC smMIPs and validated with original exome or whole-genome sequencing. Blue boxes indicate that the variant was observed on original exome or whole-genome sequencing but not identified via the CIViC smMIP capture panel. The left panel indicates the number of samples containing a mutation in the indicated gene. AML, acute myeloid leukemia; CRC, colorectal cancer; HL, Hodgkin lymphoma; HNSCC, head and neck squamous cell carcinoma; SCLC, small-cell lung cancer.

FIG 4.

Variant allele frequencies (VAFs) observed using original exome or whole-genome sequencing compared with VAFs observed using Clinical Interpretations of Variants in Cancer (CIViC) single-molecule molecular inversion probe (smMIP) sequencing. (A) Correlation of VAF with original sequencing parsed by sequencing status (ie, passed sequencing if total sequencing counts were > 1 standard deviation from the mean and tag complexity was > 600,000 unique captured smMIPs). (B) Correlation of VAF with validation status (ie, true if the variant identified using exome or genome sequencing was identified on CIViC smMIP sequencing). (C) Correlation of VAF parsed by coverage at variant loci. (D) Correlation of VAF parsed by DNA mass input for library construction. (E) Correlation of VAF parsed by presence or absence of matched normal tissue. (F) Correlation of VAF parsed by tumor type. AML, acute myeloid leukemia; CRC, colorectal cancer; HL, Hodgkin lymphoma; OSCC, oral squamous cell carcinoma; SCLC, small-cell lung cancer.

FIG 5.

Analysis of variants rescued by Clinical Interpretations of Variants in Cancer (CIViC) single-molecule molecular inversion probe (smMIP) sequencing for samples with both tumor and matched normal. There were 217 variants called as somatic by CIViC smMIP sequencing that were not identified by the original sequencing. All variants were manually reviewed using both CIViC smMIP sequencing data and original sequencing data. (A) During manual review, 55 variants were identified as germline. A histogram shows that the distribution of the smMIP variant allele frequencies (VAFs) for these germline variants was observed at 50% and 100% VAF, indicating heterozygosity and homozygosity, respectively. (B) An additional 36 variants were identified as sequencing artifacts. Most artifacts were either mononucleotide repeats (MN), dinucleotide repeats (DN), or tandem repeats (TR). Other artifacts include multiple mismatches (MM) or multiple variants (MV). (C) During manual review, 162 variants did not show any support in the original sequencing data. Most unsupported variants did not have sufficient coverage to be detected based on a binomial probability of ≤ 3 variant-supporting reads (see Materials and Methods). (D) The remaining 11 variants had variant support in original sequencing but were not called as somatic in final original annotation. The scatter plot shows correlation between original VAF and CIViC smMIP VAF for these variants.