Abstract Molecular profiling of rare cancer cells derived from liquid biopsy holds promise to monitor systemic cancer progression and guide personalized therapy interventions. However, this necessitates reliable and comprehensive profiling of samples at the single cell level and requires whole genome amplification (WGA), which adds amplification errors and bias to common sequencing errors. Therefore, we established a single-cell mutation calling workflow that can identify clinically actionable somatic mutations in rare single cell DNA with high sensitivity and accuracy. Our workflow is an end-to-end solution which can take FASTQ reads from any Panel Sequencing (PanelSeq) data and gives out single nucleotide variations (SNVs), insertion-deletions (InDels), and copy number variations (CNVs) in single cells alongside with mutation specific information from ClinVar, OncoKB, pathogenicity scores, conservations scores as well as known association to drug response and resistance. We incorporated a Bayesian Neural Network (NN) based Classifier algorithm that assigns a single statistical confidence score to the mutations, enabling prioritization of the mutations in a robust manner. On the training dataset including MDA-MB-453, BT474, BT549 and ZR-75.1 cell lines, the classifier achieved 91% sensitivity and 96% specificity (AUC=0.98), while on the independent test collective including HCC1395/HCC1395BL cell lines it achieved 84% sensitivity and 96% specificity (AUC=0.90). We applied the workflow to patient samples suffering from advanced metastatic Triple Negative Breast Cancer (TNBC) assayed with in-house single-cell Integrated Mutation Profiling of Actionable Cancer Targets (scIMPACT) and single-cell Whole Exome Sequencing (scWES). Using our method, we detected clinically actionable SNVs with pathogenic relevance in primary tumor, metastatic specimens, circulating tumor cells (CTCs) from blood and disseminated cancer cells (DCC) from pleural effusions, as well as the matched CTC-derived in vivo models (CDX). Notably, functional in vitro drug screens performed on CDX models uncovered susceptibility to approved drug treatments conferred by mutations detected in the samples. In summary, we introduce a novel mutation calling workflow facilitating detection of SNVs, Indels and CNVs in rare single cells with high sensitivity and precision. Confident scores assigned by the NN classifier enables identification of functionally actionable mutations. The workflow was executed in GDPR compliant fashion and follows all data protection standards imposed by the European laws. Accordingly, the method is well suited for molecular profiling, therapeutic target selection and longitudinal monitoring of liquid biopsy specimens in diagnostic environments. Citation Format: Adithi Ravikumar Varadarajan, Thomas Ragg, Jonas Grote, Vadim Dechand, Steffi Treitschke, Clara Chaiban, Christoph Klein, Zbigniew Czyz, Jens Warfsmann. A targeted mutation calling workflow - end to end - from FASTQ to clinical report for single cells [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2283.