Identification of Key Microbial Signatures Associated With Breast Cancer: Machine Learning-Based Approaches Using Gut...

Area of Science:

• Microbiome research
• Cancer biomarkers
• Machine learning in oncology

Background:

• Breast cancer (BC) presents significant global health challenges, necessitating novel noninvasive biomarkers, particularly for low- and middle-income countries.
• The gut microbiome's influence on host immunity and metabolism makes it a promising source for such biomarkers.
• This study investigates gut microbiome alterations in a Ghanaian cohort to identify potential BC biomarkers.

Purpose of the Study:

• To explore gut microbiome dysbiosis associated with breast cancer.
• To evaluate the efficacy of machine learning models for predicting breast cancer status using microbiome data.
• To identify specific bacterial genera correlated with breast cancer risk.

Main Methods:

• 16S rRNA gene sequencing was performed on samples from 520 BC patients and 442 healthy controls.
• QIIME2 and DADA2 were used for sequence processing, followed by diversity and community structure analyses.
• Supervised machine learning algorithms, including Random Forest, were trained and evaluated for predictive performance.

Main Results:

• Breast cancer patients exhibited reduced microbial richness and distinct gut microbiome community structures.
• Machine learning models, particularly Random Forest, achieved high predictive accuracy (AUC=0.78, accuracy=0.70) for breast cancer status.
• Specific bacterial genera, such as Bacteroides and Lachnoclostridium, were associated with elevated BC risk, while Coprococcus and Prevotella were linked to lower risk.

Conclusions:

• Significant shifts in gut microbiome composition are evident in breast cancer patients.
• Identified bacterial genera can serve as indicators of altered breast cancer risk.
• Machine learning effectively predicts breast cancer risk, paving the way for novel microbiome-based diagnostic strategies.