A NOVEL MACHINE LEARNING APPROACH FOR TUMOR DETECTION BASED ON TELOMERIC SIGNATURES

Area of Science:

• Oncology
• Genomics
• Bioinformatics

Background:

• Cancer encompasses over 200 types, each with unique molecular profiles requiring tailored therapies.
• Both short and long telomere lengths are linked to increased cancer risk, indicating telomere length variation's role in tumorigenesis.

Purpose of the Study:

• To develop and validate a machine learning model for predicting tumor status using telomere characteristics.
• To integrate telomere biology with large-scale genomic and phenotypic data for enhanced cancer diagnostics.

Main Methods:

• A supervised machine learning model was developed.
• The model was trained on telomeric read content, genomic variants, and phenotypic features.
• Data from 33 cancer types within The Cancer Genome Atlas (TCGA) program were utilized.

Main Results:

• The developed model achieved an 82.62% accuracy in predicting tumor status.
• The study highlights telomere length variation as a potential predictive biomarker in oncology.

Conclusions:

• The TeloQuest model offers a novel, multidisciplinary approach to cancer diagnostics and risk assessment.
• Integrating telomere biology with genomic and phenotypic data shows promise for improving oncological outcomes.
• The trained model is publicly available for further research and development.