Machine learning identification of molecular targets for medulloblastoma subgroups using microarray gene fingerprint...

Area of Science:

• Oncology
• Bioinformatics
• Genetics

Background:

• Medulloblastoma comprises distinct molecular subgroups (WNT, SHH, Group 3, Group 4) requiring precise classification for targeted therapy.
• Accurate molecular subtyping is crucial for understanding medulloblastoma heterogeneity and guiding treatment strategies.

Purpose of the Study:

• To introduce a structured methodology for identifying molecular targets to classify medulloblastoma subgroups.
• To develop and validate an artificial neural network (ANN) model for accurate medulloblastoma subtyping.

Main Methods:

• Utilized microarray gene expression data to train an artificial neural network (ANN) model.
• Employed Kruskal-Wallis and chi-squared tests for statistically relevant feature selection.
• Validated computational predictions using reverse transcription and digital Polymerase Chain Reaction (dPCR) on tumor samples.

Main Results:

• The ANN model achieved an average classification accuracy of 96% for medulloblastoma subgroups.
• Identified minimal gene combinations critical for distinguishing each subgroup.
• Experimental validation confirmed the computational predictions of gene expression levels.

Conclusions:

• The integration of machine learning and molecular quantification offers a reproducible framework for medulloblastoma subgroup classification.
• The proposed approach demonstrates statistical and experimental consistency for accurate subtyping.
• This methodology facilitates precise classification of medulloblastoma, supporting personalized treatment approaches.