Research Paper Volume 11, Issue 3 pp 974—985
Predicting chromosome 1p/19q codeletion by RNA expression profile: a comparison of current prediction models
- 1 Beijing Neurosurgical Institute, Capital Medical University, Beijing, China
- 2 Department of Neurosurgery, Beijing Tiantan Hospital, Capital Medical University, Beijing, China
- 3 China National Clinical Research Center for Neurological Diseases, Beijing, China
- 4 Center of Brain Tumor, Beijing Institute for Brain Disorders, Beijing, China
Received: October 20, 2018 Accepted: January 24, 2019 Published: February 2, 2019
https://doi.org/10.18632/aging.101795How to Cite
Abstract
Background: Chromosome 1p/19q codeletion is increasingly being recognized as the crucial genetic marker for glioma patients and have been included in WHO classification of glioma in 2016. Fluorescent in situ hybridization, a widely used method in detecting 1p/19q status, has some methodological limitations which might influence the clinical management for doctors. Here, we attempted to explore an RNA sequencing computational method to detect 1p/19q status.
Methods: We included 692 samples with 1p/19q status information from TCGA cohort as training set and 222 samples with 1p/19q status information from REMBRANDT cohort as validation set. We reviewed and compared five tools: TSPairs, GSVA, PAM, Caret, smoother, with respect to their accuracy, sensitivity and specificity.
Results: In TCGA cohort, the GSVA method showed the highest accuracy (98.4%) in predicting 1p/19q status (sensitivity=95.5%, specificity=99.6%) and smoother method showed the second-highest accuracy (accuracy=97.8%, sensitivity=96.4%, specificity=98.3%). While in REMBRANDT cohort, smoother method exhibited the highest accuracy (98.6%) (sensitivity= 96.7%, specificity=98.9%) in 1p/19q status prediction.
Conclusions: Our independent assessment of five tools revealed that smoother method was selected as the most stable and accurate method in predicting 1p/19q status. This method could be regarded as a potential alternative method for clinical practice in future.