Datasets:

junidude14
/

PBISC-data

paper-releases/pbisc-diffusion/updates/20261012T130738Z-strictpairs-implementation/diffusion/dataset/server_1/model8/plans/paper_figures_strictpairs_20261012/archive/20261012T130738Z-strictpairs-implementation/snapshot/manuscript/sections/abstract.tex ADDED Viewed

	@@ -0,0 +1 @@

+ Bulk RNA sequencing offers cost-effective, sensitive transcriptome profiling but averages expression across cells, obscuring cellular heterogeneity. Deconvolution methods estimate cell-type composition and expression from the bulk dataset but do not resolve variation within cell types. Existing bulk-to-single-cell approaches can be limited by the cellular diversity represented in selected references and by differences between pseudobulk and measured bulk RNA-seq. Here we introduce PBISC, a bulk-conditioned diffusion model that combines pretraining on over 33 million single cells with low-rank adaptation (LoRA) using paired bulk and single-cell data from just five donors. The adapted model generates single-cell profiles for the remaining cohort from bulk measurements alone. By repeated sampling from independent Gaussian noise, we were able to produce a synthetic cell population representing the sample's cellular heterogeneity. Pre-annotated cell types were not used for either training or generation, allowing PBISC to generate rare cell populations without defining them as separate categories in advance. Furthermore, PBISC showed improved preservation of disease-associated expression changes relative to comparator methods in most evaluated settings. Together, these results support extending analyses at single-cell resolution to large cohorts profiled by bulk RNA-seq.

	@@ -0,0 +1,26 @@

+% Declarations: data and code availability, acknowledgements, author contributions, competing interests, ethics.
+% Written 2026-10-10. Accessions checked against plans/paired_finetune_intradataset_20260927/data/<cohort>/QC_REPORT.md
+% and MANIFEST.json, the datalake GEO series titles (GSE314922 = bulk, GSE314416 = CITE-seq), the CELLxGENE source
+% citation of the AIFI atlas, plans/c1q_crossdisease_20261004/README.md (Kawasaki), the Randolph and Arunachalam source
+% records of plans/utility_studies_final6k_20260925, and the software versions in run receipts and
+% scores_contrast/*/limma/versions.txt; BayesPrism 2.2.3 from the DESCRIPTION file in the R library that
+% plans/bayesprism_paired_20261006/tools/run_bp.R loads. Author-specific statements are left as \todo.
+% Fragment only: \input by main.tex after the Methods.
+\section*{Data availability}
+All data analysed in this study are public. We used the processed expression data that the authors deposited and did not reprocess sequencing reads. The five paired cohorts are available as follows. ImMicro has bulk RNA-seq in the Gene Expression Omnibus (GEO) under GSE314922 and CITE-seq under GSE314416\cite{babdor2026immicro}. Ye 2026 sepsis has bulk RNA-seq under GSE279448 and scRNA-seq under GSE279452\cite{ye2026sepsis}; its CITE-seq series GSE279451 was not used. Imperial has bulk RNA-seq in Zenodo record 6497251 (doi:10.5281/zenodo.6497251)\cite{gisby2022multiomics} and CITE-seq gene expression in the COVID-19 Cell Atlas file \texttt{eskd\_covid19.h5ad} (\url{https://covid19.cog.sanger.ac.uk/eskd_covid19.h5ad})\cite{stephenson2025temporal}. Giroux is GEO SuperSeries GSE206284, with bulk RNA-seq under GSE206263 and scRNA-seq under GSE206283\cite{giroux2022differential}. Ji has whole-blood bulk RNA-seq under GSE287540 and scRNA-seq under GSE287541\cite{ji2025chili}. The cohort of Liu et al.\ (GSE161918)\cite{liu2021timeresolved} supplied extra bulk RNA-seq and CITE-seq training pairs, used only in Supplementary Notes~2 and 3. Its single cells are also part of the pretraining corpus, and it was never a test cohort. The lupus scRNA-seq cohort is GSE135779\cite{nehar2020mapping}. Beh\c{c}et disease bulk RNA-seq is GSE198533 and its scRNA-seq GSE198616\cite{zheng2022c1qbehcet}. Kawasaki disease scRNA-seq, for the planned Fig.~\ref{fig:fig5}e, is GSE168732\cite{wang2021kawasaki}. The two further paired cohorts of Fig.~\ref{fig:fig2}a are those of Randolph et al., with bulk RNA-seq under GSE162632 and the authors' pseudobulk inputs in Zenodo record 4273999 (doi:10.5281/zenodo.4273999)\cite{randolph2021genetic}, and of Arunachalam et al., with bulk RNA-seq under GSE152418 and scRNA-seq under GSE155673\cite{arunachalam2020systems}. The Human Immune Health Atlas of the Allen Institute for Immunology (AIFI), on which the cell-type annotator was trained, is in CZ CELLxGENE Discover (collection 77f9d7e9-5675-49c3-abed-ce02f39eef1b)\cite{gong2025multiomic}. The 22 public datasets of the pretraining corpus are summarised in Supplementary Note~1 and Supplementary Fig.~1. \todo{list the 22 corpus datasets with their accessions (manuscript source\_data/corpus/corpus\_datasets.csv, where several CELLxGENE entries lack a collection identifier) in Supplementary Note~1 or a Supplementary Table} \todo{deposit and link the pretrained model and LoRA weights, the generated cells (including the bulk-only sets of Fig.~\ref{fig:fig6}), the sample splits, conditions and scores} Source data are provided with this paper. \todo{assemble the Source Data file from the deck's data tables}
+\section*{Code availability}
+Code for pretraining, LoRA fine-tuning, sampling, scoring and every figure is available at \todo{public repository URL} and archived at \todo{Zenodo DOI of the release}. The model was trained and sampled with Python 3.12.14 and PyTorch 2.14.0. The cell-type annotator used scvi-tools 1.4.1, and the classifier tests used scikit-learn 1.9.0. The correspondence-at-the-top (CAT) analysis used R 4.5.2 with limma 3.66.0 and edgeR 4.8.1. scSemiProfiler was run from the authors' repository at commit c2aea02 with the compatibility patch described in Methods. Cell-type deconvolution used BayesPrism 2.2.3 and the official CIBERSORTx Docker images (Fractions and HiRes). \todo{full software list for the Reporting Summary}
+\section*{Acknowledgements}
+\todo{authors to add funding sources, grant numbers and the people to thank}
+\section*{Author contributions}
+\todo{authors to state who did each part of the work and to confirm that all authors approved the final manuscript}
+\section*{Competing interests}
+\todo{all authors to confirm their competing-interest statements}
+\section*{Ethics declarations}
+This study reanalysed published, de-identified expression data, and no new samples were collected from people. Approvals and consent for the original samples are those reported by the source studies. \todo{authors to confirm the reuse terms of every dataset and their institution's requirements; this draft claims no approval that has not been confirmed}

	@@ -0,0 +1,7 @@

+\section*{Discussion}
+The population-pretrained model and the exact-five-pair protocol define a practical test of whether a small number of matched donors can adapt single-cell generation to measured bulk. The Figure~\ref{fig:fig1} diagram identifies the inputs, masks and output budget, while the Figure~\ref{fig:fig2} pretraining rescore shows coverage--precision trade-offs across saved architectures and gene panels. These design results do not determine which adaptation method performs best on held-out real bulk.
+The decisive healthy, disease and rare-state comparisons remain pending under the frozen strict support manifests. The earlier five-fold analyses and rare-state caches document useful pilot behavior, but some had different validation budgets or multiple selected-donor visits. They are retained as historical evidence outside the strict estimates. The Beh\c{c}et bulk and single-cell donors are mostly unmatched, so that analysis can only be descriptive. A future complete version must report donor-level uncertainty and null comparisons alongside cell-quality, expression and rare-state measures.
+Figure~\ref{fig:fig6} awaits the user-prepared bulk-only reproduction data. This draft makes no discovery claim from those samples.

	@@ -0,0 +1,3 @@

+\mainfigure{fig:fig1}{Population pretraining, exact five-pair adaptation and measured-bulk generation}
+\figpanelh{}{figures/strict_20261012/fig1_overview.pdf}{0.77}
+\figlegend{\textbf{Overview.} The model learns from within-sample pseudobulks and adapts to measured bulk with one matched sample from each of five donors. Expression, six depth/profile scalars and two assay masks form the 18,006-entry condition. The input's nominal $K_{\mathrm{cond}}=1{,}000$ and $\rho=3{,}869.603$ counts per cell differ from the $N_{\mathrm{gen}}=1{,}500$ output cells. External AIFI reference calibration is an alternative branch without a strict-support performance claim. Bars describe all 41,395,184 corpus cells, not donors. Source: the verified Figure 1 provenance and corpus aggregate files.}