CnQuant: high-resolution chromosomal copy number profiling for precision oncology in the clinics Authors: Benjamin Maciej Freyter, Claus Hultschig, Jon Brugger, Ivana Bratic Hench, Stephan Frank, Juergen Hench Published: 2026 Combined tumor DNA methylation and chromosomal copy number profiling has transformed precision oncology during the last decade. While the standard microarray-based approach to methylome profiling facilitated the development of multiple machine learning-based tools [1, 3–5, 7–13, 15, 17], the full clinical potential of array-derived copy number (CN) data remains partially underutilized. The primary added value of CNquant lies in (1) enhanced interactive annotation of focal CNVs beyond standard static CNV profile output, and (2) cohort-level CNV frequency analyses, useful both for plausibility checks in clinical diagnostics as well as for research applications. In addition, our tool provides interactive user interfaces and extensive reference data, addresses data format incompatibilities, has low computational requirements, and is highly scalable. We tackle these shortcomings with CnQuant, an interactive Python suite (Supplement, Fig. 1) for CNV computation from DNA methylation array files with web-based visualization, which significantly extends our public "EpiDiP" platform [8]. CnQuant is free software (open-source code and locally installable in Docker/WSL) with low computational requirements and high scalability, facilitating broad clinical implementation in hospital settings. CnQuant provides harmonized, clinically oriented genomic annotation across all conventional methylation array types and versions. It comprises all required CN-neutral reference data for each covered array platform. Ad hoc generated case-specific CN plots and recurrent chromosomal aberration summaries feature a unified GUI (Supplement, Figs. 2, 3). Anonymously shareable links with preserved gene annotations enable visualization and discussion of potentially actionable targets, e.g., in the context of multidisciplinary tumor boards. Individually examinable genes aid in the interpretation of parallel sequencing variant allelic frequencies and ambiguous copy number calls.Fig. 1Overview of the CQall and CQcase graphical user interfaces. Panels (A–D) depict summary plot details as displayed by the CQall application. A Posterior fossa pilocytic astrocytomas with recurrent gain on chromosome 7 due to internal tandem duplication (LGG_PA_PF, n = 156; HM450K: 114; EPIC v1: 42). B Diffuse midline gliomas, H3K27-altered with frequent PDGFRA gain (DMG_K27, n = 89; HM450K: 78; EPIC v1: 11). C Recurrent -7/ + 10 signature of RTK II glioblastomas, IDH-wildtype (GBM_RTK_II, n = 480; HM450K: 143; EPIC v1: 249; EPIC v2: 88). D Sporadically amplified ERBB2 locus in breast carcinomas (BR_CA, n = 872; HM450K: 798; EPIC v1: 74). E–H Representative single-case plots corresponding to the tumor types in A–D. I–L Interactive tables displayed by CQall enable gene-level data exploration and contain hyperlinks to GeneCards.org. See Supplement for full GUI screenshots. For clarity, overlay summary plots were omitted; links to the respective interactive plots are provided in the Supplement Overview of the CQall and CQcase graphical user interfaces. Panels (A–D) depict summary plot details as displayed by the CQall application. A Posterior fossa pilocytic astrocytomas with recurrent gain on chromosome 7 due to internal tandem duplication (LGG_PA_PF, n = 156; HM450K: 114; EPIC v1: 42). B Diffuse midline gliomas, H3K27-altered with frequent PDGFRA gain (DMG_K27, n = 89; HM450K: 78; EPIC v1: 11). C Recurrent -7/ + 10 signature of RTK II glioblastomas, IDH-wildtype (GBM_RTK_II, n = 480; HM450K: 143; EPIC v1: 249; EPIC v2: 88). D Sporadically amplified ERBB2 locus in breast carcinomas (BR_CA, n = 872; HM450K: 798; EPIC v1: 74). E–H Representative single-case plots corresponding to the tumor types in A–D. I–L Interactive tables displayed by CQall enable gene-level data exploration and contain hyperlinks to GeneCards.org. See Supplement for full GUI screenshots. For clarity, overlay summary plots were omitted; links to the respective interactive plots are provided in the Supplement CnQuant utilizes Mepylome [2]; it comprises background services (CQmanager, CQcalc, CQall_plotter; for details on these components see Supplement) and foreground GUIs (CQall, CQcase). CQmanager coordinates data analysis from methylation array files through a local REST API facilitating integration into preexisting data workflows, and is particularly useful for the implementation of on-premise tools to generate oncology-centered reports. CQcalc computes CNVs based on gender-balanced, array-type-specific reference data and prepares per-case CNV plots. To minimize load, data are stored in a compressed, checksum-verified manner. CQall_plotter overlays and summarizes CN coordinates on intersecting genomic loci across various array types within each sample cohort. CQall and CQcase share an intuitive, unified graphical web interface for interactive data visualization and examination, eliminating the need for bioinformatics user knowledge. CQall (Fig. 1A–D) serves as an atlas. It allows clinicians to identify recurrent CNVs in reference case cohorts. For plausibility checks, usage expands into the non-microarray field, particularly the interpretation of (targeted) 2nd generation parallel and 3rd-generation nanopore sequencing results [6, 14, 16, 18]. To eliminate shortcomings associated with rare tumor entities, for which only small reference cohorts (frequently obtained with multiple array types) are available, CQall restricts analysis to overlapping probes present on all included array types. This prevents invalid, diagnostically misleading comparisons of single-case data with reference cohorts compiled from different array types. CQcase (Fig. 1E–H) lets clinicians instantly examine CNVs of each individual case. Specific genetic loci can be displayed and annotated at a high resolution. Such labeled genes may highlight the presence or absence of targetable/diagnostic gene amplifications and deletions (ERBB2, MDM2, PDGFRA; CDKN2A, etc.). Annotated plots are instantly exchangeable via web links e.g., during interdisciplinary tumor boards and are downloadable into electronic health records. In an internal validation study comprising 30 brain tumors of four types (breast cancer metastasis; diffuse midline glioma, H3K27-altered; glioblastoma, IDH-wildtype, RTK II; posterior fossa pilocytic astrocytoma) exhibiting oncologically relevant CN changes. Concordance was visually determined in 155 genomic loci. We observed a match in 153 loci (98.71%) with accredited routine diagnostic data obtained through either 2nd-generation parallel DNA/RNA sequencing (Oncomine Comprehensive Assay V2, Thermo Fisher and Archer™ FUSIONPlex™ Core Solid Tumor panel; IDTDNA, or Her2 fluorescence in situ hybridization / immunohistochemistry (Supplementary Table 1). To our knowledge, no other open-source high-resolution CNV exploration software for methylation arrays is available. None of the existing tools including Conumee 2.0 [5] and SeSAMe [19] provide interactive GUIs, on-the-fly gene annotation, high speed, and low resource allocation (see Supplement, Benchmarking chapters). In addition, CnQuant is seamlessly integrated into pan-cancer epigenetic profiling [8]. We therefore anticipate that CnQuant will significantly contribute to precision medicine by guiding personalized oncologic care decisions. Supplementary Information Below is the link to the electronic supplementary material.Supplementary file1 (PDF 812 KB)Supplementary file2 (XLSX 26 KB) Supplementary file1 (PDF 812 KB) Supplementary file2 (XLSX 26 KB)