Application of Transcriptome-Based Gene Set Featurization for Machine Learning Model to Predict the Origin of Metastatic Cancer
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.
View abstract on PubMed
Summary
This summary is machine-generated.A new AI tool, ONCOfind-AI, accurately predicts the origin of metastatic cancers using gene expression data. This approach integrates different data types, improving cancer diagnosis and treatment strategies for unknown primary cancers.
Area Of Science
- Computational biology
- Genomics
- Machine learning in oncology
Background
- Identifying the primary tumor site in cancer of unknown primary (CUP) is critical for effective treatment and patient outcomes.
- Current diagnostic methods struggle to pinpoint the origin of metastatic cancers, leading to challenges in patient management.
- CUP accounts for a significant number of cancer-related deaths, highlighting the need for improved diagnostic tools.
Purpose Of The Study
- To introduce ONCOfind-AI, a machine learning framework designed to predict the primary site of metastatic cancers.
- To leverage transcriptome-based gene set features for enhanced accuracy in cancer origin prediction.
- To demonstrate the framework's ability to integrate RNA sequencing and microarray data for improved model performance.
Main Methods
- Development of a machine learning framework (ONCOfind-AI) utilizing gene set scores derived from transcriptome data.
- Characterization of transcriptome profiles from different data platforms (RNA sequencing and microarrays) using gene set scores.
- Integration of data from diverse platforms to train and enhance machine learning models for cancer origin prediction.
Main Results
- Integration of data from different platforms significantly improved the accuracy of machine learning models for predicting cancer origins.
- External validation using clinical samples achieved a top-1 accuracy of 0.80-0.86 and a top-2 accuracy of 0.90.
- The use of curated gene sets facilitated the merging of gene expression data from disparate platforms.
Conclusions
- ONCOfind-AI demonstrates a promising approach for improving the accuracy of metastatic cancer origin prediction.
- Incorporating biological knowledge via gene sets enhances data compatibility across different platforms, crucial for robust machine learning models.
- This framework has the potential to aid in the diagnosis of CUP, guiding more effective treatment decisions and potentially improving patient outcomes.
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.
