EpiNanSarc

Background, rationale
Sarcoma subtyping is crucial for determining the prognosis of the disease, as it helps predict how rapidly the cancer will grow and spread. While several diagnostic tools exist to support subtyping and patient classification, a high inter-observer variability makes their classification harder, and 15-20% of cases remain unclassified.
Molecular analyses greatly improve classification by detecting typical genetic alterations and, recently, a methylation-based classifier has been shown to be capable of reliably classifying most sarcoma subtypes. Yet, this approach cannot be easily translated into routine diagnostics and comprehensive molecular analysis requires access to several analytical tools - not always available, especially in smaller hospitals.
Oxford Nanopore Technologies (ONT) sequencing is a competitive platform in terms of sample handling and instrumental costs. We have shown that ONT sequencing can detect multiple genome-wide genetic/epigenetic tumor DNA features from a single assay, and used tumor epigenetics to classify sarcoma.

Hypothesis
Our hypothesis is that comprehensive genome-wide characterization of sarcomas using ONT sequencing will result in enhanced diagnosis and patient stratification.

Aims
Main aim of the project is to develop a rapid, point-of-care tool to precisely diagnose sarcoma; understand sarcoma aetiology; identify novel treatment targets as well as predict and monitor treatment responses and toxicity. We will validate methylation-profiling by ONT sequencing as a viable alternative to more complex approaches. Additionally, we will develop multi-modal analysis to correlate epigenetic (methylation levels, specific alterations), genetic (point mutations, CNAs, structural variants, signatures of CNAs, mutational signatures), and clinical features with treatment outcomes and evolution of the disease.

Methods
A retrospective and a prospective study will be carried out on tumor tissues (WP1), tumor-derived cell lines (WP2), and liquid biopsy (WP3). ONT sequencing will be performed on the various samples to extract genome-wide methylation profiles, point mutations, CNAs, structural variants, signatures of CNAs, mutational signatures. Additional epigenetic/genetic features will be extracted as well. A multi-modal Machine Learning model will be trained based on all these features for sarcoma subtype classification and patient stratification.

Expected results and potential impact
We expect that our genome-wide profiling via ONT sequencing will match or improve other subtyping approaches, and that this novel approach will facilitate more effective and streamlined diagnosis and management for sarcoma patients, ultimately leading to improved outcomes.