Computers in biology and medicine · 2025
Comparative machine learning analysis for predicting organ tropism in breast cancer and identifying key gene signatures.
СтатьяСравнительное исследованиеНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Chakraborty Sohini, Shakhapur Nidhi, K Sathya, Banerjee Satarupa
Аннотация
BACKGROUND: Breast cancer metastasis (BCM) metastasizes preferentially to certain organs. Important genetic markers can be used for early detection and treatment. Machine learning (ML) can efficiently handle gene expression data to enhance metastasis prediction.
METHODS: This study employed the GSE14020 dataset of gene expression profiles of 65 breast cancer patients with liver, lung, bone, or brain metastasis, covering 22,474 genes. Feature selection by Gini reduction reduced the top 25 important genes to nine important genes. Three machine learning models-Random Forest (RF), Support Vector Machine (SVM), and Artificial Neural Networks (ANN)-were employed to predict metastatic sites from gene expression profiles. Model performance was assessed by accuracy, precision, recall, and F1-score measures. Radviz visualization investigated gene-metastasis correlations and validated the biological relevance of the identified gene markers.
RESULTS: ANN had the best accuracy (>90 %), followed by RF and SVM. Nine significant genes, i.e., CCL19, FABP7, and COL1A2, were identified as potential biomarkers. Radviz visualization confirmed gene-metastasis relationships, in agreement with biological literature. RF was effective for brain and liver metastasis but was poor for bone and lung classification.
CONCLUSION: This study is an example of the potential of ML in precision oncology through enhanced biomarker discovery and prediction of metastasis. The incorporation of ML models with biological visualization software gives insights into targeted therapy approaches in breast cancer.
METHODS: This study employed the GSE14020 dataset of gene expression profiles of 65 breast cancer patients with liver, lung, bone, or brain metastasis, covering 22,474 genes. Feature selection by Gini reduction reduced the top 25 important genes to nine important genes. Three machine learning models-Random Forest (RF), Support Vector Machine (SVM), and Artificial Neural Networks (ANN)-were employed to predict metastatic sites from gene expression profiles. Model performance was assessed by accuracy, precision, recall, and F1-score measures. Radviz visualization investigated gene-metastasis correlations and validated the biological relevance of the identified gene markers.
RESULTS: ANN had the best accuracy (>90 %), followed by RF and SVM. Nine significant genes, i.e., CCL19, FABP7, and COL1A2, were identified as potential biomarkers. Radviz visualization confirmed gene-metastasis relationships, in agreement with biological literature. RF was effective for brain and liver metastasis but was poor for bone and lung classification.
CONCLUSION: This study is an example of the potential of ML in precision oncology through enhanced biomarker discovery and prediction of metastasis. The incorporation of ML models with biological visualization software gives insights into targeted therapy approaches in breast cancer.