Background This is Zenodo version 2.0 of the ITLAS HBV immune-tolerant phase reanalysis, corresponding to analysis version V19. This record supersedes the previous V18-v2 release under the same Zenodo version chain. This study is a donor-level, tissue-separated single-cell transcriptomic reanalysis of GSE182159 originally published by Zhang et al. (Gut, 2023). The dataset comprises approximately 243,000 immune cells from 23 donors, including normal controls (NL), immune-tolerant chronic hepatitis B (IT), immune-active chronic hepatitis B (IA), chronic resolved (CR), and an independent acute resolved (AR) comparator cohort. The analysis tests whether the immune-tolerant phase is transcriptionally quiescent or instead shows immune engagement without canonical effector conversion. The findings are presented as hypothesis-generating observations consistent with effector restraint and require future functional and protein-level validation. Key features of V19 Donor-level and tissue-separated framework.All comparisons were performed at the patient level, with cell-level expression aggregated to donor-level values. Liver and blood compartments were analyzed separately to avoid tissue-composition artifacts. Multiple-testing correction used stratified Benjamini–Hochberg FDR, with the primary Level B stratum defined as comparison × tissue × lineage. Tier 1 / Tier 2 evidence reporting.Findings are explicitly classified by evidence strength. Tier 1 denotes findings surviving stratified BH FDR correction at q < 0.10. Tier 2 denotes nominal p < 0.05 with donor-pair consistency ≥80% but without FDR survival. All reported gene-level findings in V19 are Tier 2; Tier 1 designation is reserved for pathway-level findings that survive FDR correction. Six concurrent transcriptomic patterns.V19 prioritizes six transcriptomic patterns consistent with effector restraint in the IT phase:P1, myeloid immunoregulatory programming;P2, epigenetic-regulator upregulation;P3, metabolic–effector dissociation;P4, JAK1–SOCS discordance;P5, liver-dominant T-cell exhaustion programming; andP6, impaired terminal differentiation signature. Of 33 prioritized gene–lineage–tissue combinations, 32 were classified as ROBUST or VALID by Oaxaca–Blinder decomposition, supporting predominantly per-cell rather than composition-driven effects. BCR/TCR repertoire integration.V19 includes donor-level BCR/TCR repertoire analysis using patient-ID grouping. The IT phase showed a BCR-dominant pattern characterized by class-switching features and minimal non-singleton emergence, consistent with a “switched but not productively expanded” BCR phenotype. TCR clonality did not show significant NL→IT expansion. Hypothesis-generating interpretation.All gene- and pattern-level claims are framed as transcriptomic observations consistent with effector restraint. The study does not claim causal mechanism, antigen specificity, or functional validation. Methods summary The analysis used donor-level Mann–Whitney U testing with stratified Benjamini–Hochberg FDR correction. Six major immune lineages were analyzed across two tissue compartments; γδ T cells were excluded due to insufficient representation. Pathway scoring used AUCell-based gene-set scoring. Gene-to-gene correlation analysis used donor-level Spearman correlations, with the Figure 4 display network restricted to strong correlations, |ρ| ≥ 0.7, and stratified BH q < 0.10. Data and code Source dataset: NCBI GEO GSE182159Original study: Zhang et al., Gut, 2023Code, notebooks, and intermediate outputs: https://github.com/choccoba/ITLAS Status This is a computational reanalysis. The results should be interpreted as hypothesis-generating. Functional assays, protein-level validation, antigen-specific receptor analysis, and independent cohort validation are not included in this release and will be required to confirm the proposed biological patterns. Relationship to previous version V19 / Zenodo version 2.0 supersedes V18-v2 / Zenodo version 1.x. Substantive revisions include: refined Tier 1 / Tier 2 evidence framework; donor-level BCR/TCR repertoire recalculation using patient-collapsed grouping; correction of repertoire metric definitions, including clone-level singleton reporting; clearer separation of gene-level nominal findings from FDR-surviving pathway-level findings; revised tissue-separated visualization and statistical reporting; more conservative “consistent with effector restraint” language throughout. The previous version remains accessible through its version-specific Zenodo DOI for citation continuity. Readers should cite V19 / Zenodo version 2.0 for the current results.
Young Min Park (Tue,) studied this question.