Evaluation of the performance of promising host transcriptional biosignatures in predicting TB progression during a 1-year prospective cohort study in Indian household contacts.

DOI of corresponing publication in Frontiers in Tuberculosis: 10.3389/ftubr.2026.1750007 The data consits of rawdata on gene expression by clinical groups; TB progressors and non-progressors. Gene expression is measured on the Fluidigm platform, expressed as ct-values, at the time of TB contact investigation in Indian households during a prospective 1-year follow-up for progression to TB. Of clincial data, gender and age groups are available. Of methodological data, quality controls and meltingpoint thresholds are given. Variable coding in the dataset:SubjectIDanonymous: Anomynized, randomly assigned number 1-192. Original PID deleted.Gender: 0=male; 1=femaleAge_strata: 1= <5years; 2= 5-14 years; 3= >14yearsFluidigm ct value: The upper cycle-limit for the Fluidigm assay: 30 cycles. Thus samples with >=cycles are given the value 999Call: PASS= The reaction meets the software’s quality‑control criteria. The amplification curve, baseline, noise level, and signal fall within acceptable limits. The Cq/Ct value is considered reliable. FLAG= No amplification (NOAMP), High noise (NOISE flag), Failed baseline/threshold detection (CTFAIL, THOLDFAIL (threshold fail)) Off‑scale signal, Abnormal melt curve (MTP) or Low signal or passive reference issues (LOWROX, BADROX)