Multiclonal Profiling of FLT3-ITD in AML Using MinION Sequencing: A Tailored Clustering Approach to Enhance Subclonal Detection

Jordi Martínez-Serra,1 Aser Alonso-Carballo,1 Ángel Horrillo,1 Paula Gómez,1 Oliver Vögler,2 Antonio Gutiérrez,1 Antonia Sampol1 1Hematology Department, Son Espases University Hospital, IdISBa, Palma, Spain; 2Group of Advanced Therapies and Biomarkers in Clinical Oncology, IUNICS, University of the Balearic Islands, Palma, SpainCorrespondence: Jordi Martínez-Serra, Molecular Diagnostic Laboratory, Son Espases University Hospital, Palma de Mallorca, Spain, Email [email protected]: Internal tandem duplications (FLT3-ITDs) in the FLT3 gene constitute a key driver mutation in acute myeloid leukemia (AML), strongly associated with poor prognosis and therapeutic resistance. Although general-purpose structural variant callers such as Sniffles have been used to detect FLT3-ITDs, their limitations in resolving clonal diversity and low-frequency variants can lead to underrepresentation of minor clones. These shortcomings highlight the need for a dedicated bioinformatics pipeline.Materials and Methods: We developed a custom clustering-based pipeline to overcome the constraints of generic SV callers, leveraging Oxford Nanopore’s MinION for sequencing. Our method focuses on FLT3-ITDs by grouping near-identical insertions into biologically meaningful subclones, thereby allowing accurate variant detection of even low-frequency events. The pipeline was benchmarked against capillary electrophoresis (CE) and Sniffles at various thresholds (including 10%, 20%, and 50% allele-frequency cutoffs), with results validated via IGV inspection and cross-mapping.Results: The pipeline successfully detected FLT3-ITDs across all tested samples, including low-frequency variants and diverse subclones that Sniffles overlooked. Analyses uncovered complex multiclonal architectures composed of dominant clones (~20– 25% of reads) plus multiple minor subclones differing in length, sequence, and breakpoint. Crucially, our approach identified duplications as short as 15 bp—events often dismissed by conventional SV callers. Comparative analyses showed that Sniffles failed to call several biologically validated ITDs detected by our custom pipeline.Conclusion: Third-generation sequencing combined with a tailored clustering strategy enhances the detection of FLT3-ITDs and clonal diversity in AML compared to generic variant callers. This method provides critical insights into subclonal populations driving relapse and therapeutic resistance—particularly in relapsed/refractory AML—underscoring the importance of specialized pipelines for precision medicine in leukemia.Keywords: acute myeloid leukemia, AML, third-generation sequencing, TGS, FLT3-ITD, minION