This archive contains the complete analysis notebook (canlung_dissociation_final.ipynb) accompanying Jacobs, Detectable and Lethal Lung Cancer Are the Same Proteomic Disease (2026), the third study in a trilogy examining whether the circulating proteins that predict cancer detection differ from those that predict cancer death (prostate: complete dissociation; colorectal: partial; lung, this work: none). The notebook reproduces every number in the manuscript from UK Biobank source data (Application 596880): a proteome-wide screen of 2,922 Olink Explore 3072 proteins against incident lung cancer (n = 544) and lung-cancer–specific mortality (n = 454) in 52,949 participants, endpoint-specific panel construction, and a pre-registered validation battery — smoking adjustment, never-smoker stratification, a neutral-dosimeter falsification with orthogonalization against the cancer axis, a second-layer screen beneath the dominant proteomic axis, a physiological confounder gauntlet (FEV1/FVC, BMI, renal function, prevalent COPD/asthma), and diagnosis-landmark and far-prodrome analyses. The archive includes all screen exports (primary, smoking-adjusted, never-smoker, second-layer, fully adjusted), top-25 tables with bootstrap confidence intervals, full logistic regression coefficients for both portable panels, and figure sources. UK Biobank data cannot be redistributed; the notebook runs against locally held parquet extracts available to approved researchers via https://www.ukbiobank.ac.uk. Analyses use folded concordance (C*), L2-regularized logistic regression at natural prevalence, 5-fold out-of-fold evaluation, and paired bootstrap (2,000 resamples); random seed, censoring date, and package versions are printed by the notebook's reproducibility cell.
Laurence A Jacobs· Zenodo (CERN European Organi...· 0 citations
This archive contains the complete analysis notebook (canlung_dissociation_final.ipynb) accompanying Jacobs, Detectable and Lethal Lung Cancer Are the Same Proteomic Disease (2026), the third study in a trilogy examining whether the circulating proteins that predict cancer detection differ from those that predict cancer death (prostate: complete dissociation; colorectal: partial; lung, this work: none). The notebook reproduces every number in the manuscript from UK Biobank source data (Application 596880): a proteome-wide screen of 2,922 Olink Explore 3072 proteins against incident lung cancer (n = 544) and lung-cancer–specific mortality (n = 454) in 52,949 participants, endpoint-specific panel construction, and a pre-registered validation battery — smoking adjustment, never-smoker stratification, a neutral-dosimeter falsification with orthogonalization against the cancer axis, a second-layer screen beneath the dominant proteomic axis, a physiological confounder gauntlet (FEV1/FVC, BMI, renal function, prevalent COPD/asthma), and diagnosis-landmark and far-prodrome analyses. The archive includes all screen exports (primary, smoking-adjusted, never-smoker, second-layer, fully adjusted), top-25 tables with bootstrap confidence intervals, full logistic regression coefficients for both portable panels, and figure sources. UK Biobank data cannot be redistributed; the notebook runs against locally held parquet extracts available to approved researchers via https://www.ukbiobank.ac.uk. Analyses use folded concordance (C*), L2-regularized logistic regression at natural prevalence, 5-fold out-of-fold evaluation, and paired bootstrap (2,000 resamples); random seed, censoring date, and package versions are printed by the notebook's reproducibility cell.
Laurence A Jacobs· Zenodo (CERN European Organi...· 0 citations
This archive contains the complete analysis notebook (canlung_dissociation_final.ipynb) accompanying Jacobs, Detectable and Lethal Lung Cancer Are the Same Proteomic Disease (2026), the third study in a trilogy examining whether the circulating proteins that predict cancer detection differ from those that predict cancer death (prostate: complete dissociation; colorectal: partial; lung, this work: none). The notebook reproduces every number in the manuscript from UK Biobank source data (Application 596880): a proteome-wide screen of 2,922 Olink Explore 3072 proteins against incident lung cancer (n = 544) and lung-cancer–specific mortality (n = 454) in 52,949 participants, endpoint-specific panel construction, and a pre-registered validation battery — smoking adjustment, never-smoker stratification, a neutral-dosimeter falsification with orthogonalization against the cancer axis, a second-layer screen beneath the dominant proteomic axis, a physiological confounder gauntlet (FEV1/FVC, BMI, renal function, prevalent COPD/asthma), and diagnosis-landmark and far-prodrome analyses. The archive includes all screen exports (primary, smoking-adjusted, never-smoker, second-layer, fully adjusted), top-25 tables with bootstrap confidence intervals, full logistic regression coefficients for both portable panels, and figure sources. UK Biobank data cannot be redistributed; the notebook runs against locally held parquet extracts available to approved researchers via https://www.ukbiobank.ac.uk. Analyses use folded concordance (C*), L2-regularized logistic regression at natural prevalence, 5-fold out-of-fold evaluation, and paired bootstrap (2,000 resamples); random seed, censoring date, and package versions are printed by the notebook's reproducibility cell.
Laurence A Jacobs· Zenodo (CERN European Organi...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
