Deep learning enabled integration of tumor microenvironment microbial profiles and host gene expressions for interpretable survival subtyping in diverse types of cancers
2024

Deep Learning for Cancer Survival Subtyping

Sample size: 20 publication 10 minutes Evidence: high

Author Information

Author(s): Zhang Haohong, Xiong Xinghao, Cheng Mingyue, Ji Lei, Ning Kang

Primary Institution: Huazhong University of Science and Technology

Hypothesis

Can integrating tumor microbiome and host gene expression data improve cancer survival subtyping?

Conclusion

The ASD-cancer framework successfully identifies distinct survival subtypes across 20 cancer types, enhancing risk stratification and biomarker discovery.

Supporting Evidence

  • ASD-cancer identified two distinct survival subtypes across all 20 types of cancer.
  • The framework provided improved risk stratification compared to traditional methods.
  • Potential interactions between microbes and host genes were identified that may influence survival.
  • Validation on external cohorts demonstrated the model's predictive capabilities.

Takeaway

This study uses a smart computer program to look at tiny living things in tumors and genes to help doctors figure out how long patients might live with cancer.

Methodology

The study developed a deep learning framework called ASD-cancer to analyze RNA-seq and tumor microbiome data from The Cancer Genome Atlas.

Limitations

The study primarily used data from Western populations, which may limit generalizability to other populations.

Participant Demographics

The study analyzed data from 20 types of cancer represented in The Cancer Genome Atlas.

Statistical Information

P-Value

8.12E−6

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1128/msystems.01395-24

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication