GDIS-Bio v1.0.0 – Manuscript Reproducibility Release

This is the first reproducibility release of GDIS-Bio, associated with the manuscript:

“GDIS-Bio: A Generalized Dynamical Instability Framework for Localizing Transcriptional State Transitions in Single-Cell Trajectories”

Authors: Hamid Ismail, Ahmed Harb, Basem William, and Marwan Bikdash

This release contains the complete computational workflow used to reproduce the analyses reported in the study, including:

public-data acquisition from NCBI GEO;
dataset qualification and cohort definition;
single-cell preprocessing and transcriptional state-space construction;
pseudotemporal trajectory validation and sliding-window generation;
GDIS calculation using pyGDIS v1.0.0;
state-space dimensionality and window-size sensitivity analyses;
conventional transition-associated metric benchmarking;
structure-preserving circular-shift null analyses;
external validation in an independent cardiac differentiation dataset;
paired statistical comparisons; and
generation of manuscript figures, supplementary figures, and summary tables.

The workflow analyzes two public single-cell RNA-seq datasets:

GSE114412 — human pancreatic endocrine differentiation;
GSE175634 — human iPSC-derived cardiac differentiation.

The repository includes the sequential analysis scripts, environment specifications, frozen analysis parameters, Linux and Windows execution scripts, selected reference outputs, manuscript assets, and documentation linking the reported results to the corresponding computational stages.

Large raw and intermediate data files are not included in the release. They can be downloaded or regenerated using the supplied scripts.

Repository: https://github.com/hamiddi/GDIS-Bio

Version: v1.0.0

License: MIT