Example datasets

An staver workflow can be divided into the following two stages. We provide separate tutorials for each stage, using the quality control data as an example.

We provide a tutorial for running the STAVER workflow on DIA-MS dataset with the Command-Line Interface (CLI) code.

To evaluated the robust generalization performance of the STAVER algorithm across diverse and large-scale DIA datasets, we applied STAVER to a much more diverse and larger-scale DIA dataset from the “ProCan-DepMap-Sanger project” (Gonçalves, et al. 2022, Cancer Cell). We provide the comprehensive analysis of the PanCancer DIA-MS dataset, complete with detailed steps and corresponding code.

• Execute the STAVER workflow
  • List all the optional command arguments
  • Run the staver_pipeline
• STAVER shows robust generalization across various large-scale DIA datasets
  • Metadata
  • Tissue_type
  • Cancer_type
  • Cancer_subtype
  • The Sankey diagram delineates the relationships
  • The reproducibility and reliability of the STAVER-processed data
    • HEK293T Spearman correlation analysis
    • The correlation heatmap of raw data
    • The correlation heatmap of STAVER-processed data
    • The IQR of HEK293T spearman correlation matrix in rawdata
    • The IQR of HEK293T spearman correlation matrix in SATVER-processed data
    • The identification protein numbers of the raw data and STAVER data
    • The Coefficient of Variation (CVs) of the raw data and STAVER data
  • The advantages of the STAVER algorithm to uncover inherent biological differences
    • The UMAP analysis of diverse 1242 cancer cell line samples
    • The UMAP analysis of the Rawdata
    • The UMAP analysis of STAVER-processed data
    • Cancer specific proteins
    • CTSE overexpression in Gastric, pancreatic, and colorectal cancer of HPA dataset
    • GPA33 overexpression in colorectal cancer and Gastric cancer of HPA dataset
    • ADGRF1 overexpression in pancreatic cancer of HPA dataset
  • The reproducibility of the STAVER algorithm in identifying the previously reported tumor biomarkers
  • The robustness and broad applicability of the STAVER algorithm for disease diagnosis and classification.
• STAVER’s robustness and applicability in disease diagnosis and classification.
  • Define the DeLong’s test function and test it
  • Define the benchmark models
  • Define the benchmark models with 95% CI AUC scores
  • Load the Original data and the STAVER-processed data
  • Train the benchmark models based on the reported cancer biomarkers
    • Hepatocellular Carcinoma biomarkers
    • Gastric Carcinoma biomarkers
    • Colorectal Carcinoma biomarkers
    • NSCLC vs. SCLC (NSCLC and SCLC biomarkers)
    • Glioma biomarkers
    • Pancreatic Carcinoma biomarkers
  • kindney Carcinoma
  • The summary of the benchmark models’ performance acorss 7 cancer types