iProMix: A mixture model for studying the function of ACE2 based on bulk proteogenomic data
Xiaoyu Song1, Jiayi Ji2, Pei Wang3
1Tisch Cancer Institute, Institute for Health Care Delivery Science, Department of Population Health Science and Policy, Icahn School of Medicine at Mount Sinai, New York, NY.
Journal of the American Statistical Association
|July 6, 2023
Summary
A new statistical framework, iProMix, identifies epithelial cell protein associations with ACE2, revealing sex-specific interferon pathways linked to COVID-19 severity. This aids understanding of sex differences in disease outcomes.
Area of Science:
- Computational Biology
- Proteomics
- Immunology
Background:
- Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) utilizes the ACE2 protein for human cell entry, necessitating research into ACE2-interacting proteins and pathways.
- Current proteomic profiling lacks single-cell resolution, hindering the examination of protein activities in specific disease-relevant cell types.
- Understanding ACE2 interactions is crucial for developing targeted therapies against COVID-19.
Purpose of the Study:
- To develop and validate a novel statistical framework, iProMix, for identifying epithelial-cell specific protein and pathway associations with ACE2 using bulk proteomic data.
- To analyze lung adenocarcinoma proteomic data to uncover ACE2-associated pathways in epithelial cells.
- To investigate potential sex-specific differences in these associations.
Main Methods:
- Proposed iProMix, a statistical framework employing a mixture model to decompose bulk proteomic data and model cell-type-specific conditional joint protein distributions.
- Enhanced cell-type composition estimation and utilized a non-parametric inference framework to address uncertainty in cell-type proportion estimates during hypothesis testing.
- Applied iProMix to proteomic data from 110 normal lung tissue samples (Clinical Proteomic Tumor Analysis Consortium).
Main Results:
- Simulations confirmed iProMix's well-controlled false discovery rates and favorable statistical power.
- Analysis of lung tissue proteomic data identified interferon alpha/gamma response pathways as significantly associated with ACE2 protein abundance in epithelial cells.
- The direction of the association between ACE2 and interferon pathways was found to be sex-specific.
Conclusions:
- iProMix is an effective statistical framework for uncovering cell-type-specific protein associations from bulk proteomic data.
- The identified sex-specific association between ACE2 and interferon pathways provides insights into the differential incidence and outcomes of COVID-19 in males and females.
- Findings suggest the need for sex-specific evaluation of interferon-based therapies for COVID-19.
