sarhnagniafariba-maker/escc-ml-biomarker-discovery

How sarhnagniafariba-maker/escc-ml-biomarker-discovery works

This repository is a research analysis for identifying ESCC gene signatures and classifying tumor versus normal samples. The documented workflow applies preprocessing, variance filtering, ANOVA feature selection, and standardization before evaluating Logistic Regression and Random Forest with LOOCV; metrics and permutation testing assess performance, while visualizations support interpretation. The README also identifies GEO GSE17351 and two data files. No notebook source was available, so edges reflect documented relationships only; the detailed implementation and exact data-file roles cannot be verified.

Main entry points: REduced_featres_conda.ipynb Analysis notebook · GSE17351_series_matrix.txt GSE17351 matrix · expression_matrix.csv Expression matrix

Components 19

Expression data

GSE17351 matrix Data/GSE17351_series_matrix.txt
Expression matrix Data/expression_matrix.csv

Feature preparation

Analysis notebook REduced_featres_conda.ipynb
Preprocessing REduced_featres_conda.ipynb
Variance filtering REduced_featres_conda.ipynb
ANOVA selection REduced_featres_conda.ipynb
Standardization REduced_featres_conda.ipynb

Classification

LOOCV REduced_featres_conda.ipynb
L1 Logistic Regression REduced_featres_conda.ipynb
Random Forest REduced_featres_conda.ipynb

Evaluation

Performance metrics REduced_featres_conda.ipynb
Permutation testing REduced_featres_conda.ipynb

Analysis outputs

PCA REduced_featres_conda.ipynb
Expression heatmap REduced_featres_conda.ipynb
ROC curves REduced_featres_conda.ipynb
Confusion matrices REduced_featres_conda.ipynb
Feature importance REduced_featres_conda.ipynb
Stable gene signatures REduced_featres_conda.ipynb

Other components

GEO dataset

Connections 19

GEO dataset → GSE17351 matrix: provides data README.md
GSE17351 matrix → Analysis notebook: analyzed in README.md
Expression matrix → Analysis notebook: analyzed in README.md
Analysis notebook → Preprocessing: performs README.md
Preprocessing → Variance filtering: precedes README.md
Variance filtering → ANOVA selection: precedes README.md
ANOVA selection → Standardization: precedes README.md
Standardization → LOOCV: evaluated with README.md
LOOCV → L1 Logistic Regression: evaluates README.md
LOOCV → Random Forest: evaluates README.md
L1 Logistic Regression → Performance metrics: measured by README.md
Random Forest → Performance metrics: measured by README.md
Performance metrics → Permutation testing: assessed by README.md
Analysis notebook → PCA: visualizes with README.md
Analysis notebook → Expression heatmap: visualizes with README.md
Performance metrics → ROC curves: visualizes with README.md
Performance metrics → Confusion matrices: visualizes with README.md
Analysis notebook → Feature importance: analyzes README.md
Feature importance → Stable gene signatures: supports interpretation of README.md

Drawn by GitDiagram on November 13, 2026 from the repository's file tree, README and source files. Read as Markdown ·