NFKB–P53 Rheostat Analysis Pipeline (R)

Abstract

NFKB–P53 Rheostat Analysis Pipeline (R) This record accompanies version 1.0 of the NFKB–P53 Rheostat Analysis Pipeline, an open-source R implementation of the rheostat model described in "The NFKB-P53 antagonism as part of a rheostat network: practical implication in solving the leukemia relapse principle" (Spiros Vlahopoulos, ORCID 0000-0001-6245-3863). The pipeline operationalizes the central abstraction of that work — that the antagonism between the inflammatory transcription factor complex NFKB and the tumor suppressor P53 can be collapsed into a single interpretable axis — as executable research code. It implements all eleven tasks of the article's proposed outline: definition of curated NFKB-activated and P53-activated gene sets (with helpers to fetch full MSigDB Hallmark sets), ingestion of bulk or single-cell RNA-seq data, per-sample scoring of both programs via ssGSEA or a transparent mean-of-z method, computation of the rheostat value (Rheostat = NFKB score − P53 score), visualization of sample positions along the axis, statistical comparison across clinical states such as diagnosis, remission, and relapse, association testing with survival and minimal residual disease endpoints, robustness checks across upstream drivers (mutations, infections, radiation, inflammatory triggers), optional integration of multi-omic layers such as ATAC-seq and proteomics, and derivation and evaluation of a simple high-risk decision rule. A complete simulated leukemia cohort allows the entire pipeline to run end-to-end without any data, so that users can verify their installation and learn the workflow before substituting their own datasets. The release consists of the pipeline script together with a beginner's guide that requires no prior R experience, covering installation, execution, adaptation to real data, and troubleshooting. The code is provided for research and exploratory purposes only: the rheostat is a functional readout intended to support biological insight and hypothesis generation, it summarizes downstream state and does not itself establish causality, and it is not intended to replace clinical judgment or to serve as the sole basis for diagnosis or treatment decisions. Any clinical application requires rigorous external validation, prospective evaluation, and regulatory review appropriate to the intended context of use. The gene lists shipped by default are indicative subsets and should be replaced with fully curated, published gene sets for any real analysis. This software and its accompanying documentation were prepared with the assistance of Mistral AI (Vibe) and are released under the MIT License; users are encouraged to cite this record when reusing or adapting the pipeline, and to report issues or extensions via the associated repository.