Clustering Digestive Tract Tumors Using Transcriptomic and Mutation Data.

Area of Science:

• Genomics
• Bioinformatics
• Cancer Research

Background:

• Digestive tract cancers are typically classified by their tissue of origin.
• Transcriptome-based molecular clustering often yields similar classifications.
• Existing methods may not fully capture distinct molecular subtypes.

Purpose of the Study:

• To introduce Newmanization, a novel method for reducing tissue-specific signals in transcriptomic analysis.
• To evaluate the effectiveness of Newmanized data in classifying digestive tract cancers.
• To compare Newmanization with traditional mutation data for cancer subtyping.

Main Methods:

• Developed the Newmanization technique to refine transcriptomic data.
• Analyzed 1635 digestive tract cancer samples from The Cancer Genome Atlas.
• Utilized RNA-Seq and whole exome sequencing data.
• Compared Newmanized transcriptome and mutation data using silhouette widths and dimension reduction plots.

Main Results:

• Newmanized transcriptome data demonstrated clearer separation and higher average silhouette widths compared to unadjusted data.
• Newmanized clusters revealed a higher frequency of specific messenger RNAs (mRNAs).
• Clusters derived from Newmanized data showed distinct molecular features.

Conclusions:

• The Newmanization method significantly improves the molecular classification of digestive tract cancers.
• Newmanized data provides a more refined basis for identifying cancer subtypes.
• This approach shows promise for advancing personalized transcriptomic medicine.