Extract figures, tables, methods, and underlying data to audit results.
Press Enter β΅ to review
Explore by Goal
"In the fields of observation chance favors only the prepared mind."
- Louis Pasteur
Quick Explanation
Copied
Paper in one line (what it does)
Builds a pig (Sus scrofa) immune single-cell atlas spanning primary & secondary lymphoid organs, then provides comparative pigβhuman mapping and an interactive querying app (Shiny-PIGGI) for immune cell exploration.
Long Explanation
A single-cell immune atlas of primary and secondary lymphoid organs in pigs β rigorous visual review
Evidence base is restricted to what the provided paper text explicitly states. All quantitative summaries below are extracted from the paperβs reported cell counts.
Core claim being tested
That a pig immune atlas across bone marrow, thymus, spleen, lymph node can provide reliable immune cell identities, pigβhuman similarities, and a usable non-computational query interface.
Study scope (as stated)
Two adult (~6-month-old) intact male Yorkshire pigs; tissues: bone marrow, thymus, ileocecal lymph node, spleen; 10X 3β² libraries (v2/v3 chemistry); sequencing: HiSeq3000, 100 bp PE.
1) Data coverage (what you actually have)
Primary visual summary of the reported post-QC cell counts per tissue.
Cell counts used above are the paperβs reported final datasets after QC: 5,899 bone marrow; 17,940 thymus; 20,210 lymph node; 5,621 spleen; total 49,670 cells.
2) Methods: pipeline choices that can strongly affect cell identity calls
What the paper says it did (reproducible audit trail)
Input genome + initial processing. Reads aligned to Sus scrofa 11.1 + v97 annotation with Cell Ranger v4.0; ambient RNA handled with SoupX; filtering thresholds applied (min genes, min UMIs, max mitochondrial fraction) and doublets removed with Scrublet.
Normalization + integration + clustering. Seurat v4.3.0.1 with SCT-based integration across tissues; PCA used for downstream reductions and clustering.
Annotation strategy. Cluster marker assessment (canonical markers + DE genes). Special handling: bone-marrow cluster with mixed progenitors/B/myeloid could not be cleanly separated by increasing resolution, so it was reprocessed as a subset; clusters with high hemoglobin expression and conflicting profiles were removed as likely technical aggregates.
Cross-species mapping. Reference mapping uses porcine thymus and human spleen references; for human mapping the paper βhumanizesβ pig data prior to Seurat-based mapping with CCA anchors.
Interaction inference and spatial reconstruction. CellChat v2.1.2 (ligand-receptor using a human database after pig βhumanizationβ for ligand/receptor genes) and CSOmapR for 3D organization reconstruction based on ligand-receptor mediated self-assembly.
Interface. Shiny app βShiny-PIGGIβ built in R to allow interactive exploration: gene expression queries, differential expression (Wilcoxon), and reference mapping visualization; downloadable formats include .cloupe and .h5seurat.
3) Key biological results (only what is explicitly supported by the paper text you provided)
3.1 Cell-type breadth across all four lymphoid organs
The paper reports scRNA-seq identification of immune lineages across tissues: progenitors in primary organs; and T/ILC, B/ASC, and myeloid lineages identified in each tissue based on canonical marker gene expression and differential expression.
The paper claims that pig splenic ILC-like populations map to human NK subsets (e.g., cytotoxic ILCs predicted as human FCGR3A+ NK cells; NCR1+EOMES+ ILCs predicted as CD160+ NK cells) with hierarchical clustering and 2D embedding supporting pigβhuman intermixing when pig and human subsets are integrated.
3.3 Lymph node structure and germinal-center-associated signaling inference
Using CellChat + CSOmapR on humanized lymph node data, the paper reports a reconstructed βfollicle-likeβ 3D organization where the center contains the highest density and the largest numbers of B cells, with KLF2β T/ILC and follicular CD4+ T cells positioned closer to the structure center and associated with more significant interactions. It also reports strongest inferred pathways involving germinal-center immune processes and identifies monocyte/MΞ¦/cDCs, CD4 ab T cells, and B cells as key senders/receivers for many pathways.
3.4 In situ staining used as an anchor for gd T localization
The paper uses in situ TRDC RNA detection (RNAscope) and CD3e immunohistochemistry to support its claim that Ξ³Ξ΄ T cells localize outside germinal centers in pig lymph node and contribute minimally to inferred germinal-center immune induction, consistent with the ligand/receptor reconstruction outputs.
4) Skeptical critique: where the atlas might mislead (known unknowns)
4.1 Biggest statistical limitation: n=2 animals
The paper explicitly states its datasets are from two animals and not representative of full pig life stages, gender, rearing environment, disease status, or genetic predispositions. This limits how strongly you can generalize βcell identitiesβ as universal baseline pig immunotypes.
4.2 Annotation bias from reference mapping & βhumanizationβ
Cross-species conclusions depend on reference-based mapping and on translating porcine ligand/receptor information to a human CellChat interaction database (via βhumanizedβ gene sets). That can inflate apparent conservation for pathways that are present in the reference DB and dampen pathways that arenβt, so mapping-based similarity should be treated as evidence for transcriptional resemblance rather than proof of conserved functional circuitry.
The paper reports stromal/epithelial cells were not identified, which it attributes to leukocyte-oriented isolation and lack of enzymatic tissue digestion, and it also notes cryopreservation protocols as a contributing factor. This matters because germinal center organization and ligandβreceptor signaling often involve stromal/follicular niches; missing those compartments can bias inferred interaction networks toward leukocyte-only circuitry.
5) Reproducibility & data accessibility (what you can actually download)
All links above are listed in the paperβs Data availability statement included in your provided text.
Next best user actions (BGPT navigation)
Author reviews (direct to BGPT)
Feedback:
Updated: April 29, 2026
BGPT Paper Review
Study Novelty
80%
While pig scRNA-seq atlases exist, the paperβs novelty is its immune organβcentric coverage across primary + secondary lymphoid organs in pigs, plus pigβhuman mapping and an explicit interactive query app (Shiny-PIGGI) positioned for non-computational exploration.
Scientific Quality
70%
Strengths include a detailed stated QC/mapping workflow, multiple computational modules (integration, reference mapping, CellChat, CSOmapR), and at least one in situ validation anchor for Ξ³Ξ΄ T localization. Main quality concerns are the very small animal number (n=2), probable compartment dropout for stromal/epithelial due to leukocyte-focused dissociation, and reliance on cross-species βhumanizationβ/reference-based inference for interaction networks.
Study Generality
70%
It improves general immune-cell annotation for pigs specifically in lymphoid organs and provides a pigβhuman transcriptional similarity framework, but generalization to other pig breeds, ages, disease states, and mucosal compartments is limited by the narrow sampling.
Study Usefulness
90%
The atlas appears immediately useful as a baseline for pig lymphoid immune cell identities and as a tooling resource: publicly deposited raw data (ENA), scripts, and an interactive Shiny app supporting gene queries and reference mapping.
Study Reproducibility
70%
Reproducibility is supported by code availability and raw data deposition, but the paper does not guarantee complete reproduction solely from the provided text (and it involves multiple choices: QC thresholds, integration, mapping humanization, ligandβreceptor databases, and interaction reconstruction). The small animal n also limits statistical reproducibility of fine-grained subsets.
Explanatory Depth
70%
The paper provides strong descriptive explanations (cell-type annotation logic, pigβhuman mapping outcomes, and inferred lymph node organizational gradients) but functional causality is not directly tested beyond limited in situ localization. The inferred interaction networks are model-based and depend on ligandβreceptor resources rather than direct perturbation.
It will load the atlas-derived cell-count summaries and generate publication-ready QC/coverage plots (bar + pie) for each lymphoid organ, using the reported post-QC cell numbers.
Get emailed when your analysis is done!
We'll email you the results when your analysis is finished.
Hypothesis Graveyard
Strongman: gd T cells are major drivers of germinal-center induction in pig lymph nodes. Why itβs less favored: the paper uses both scRNA-seq-based inference and in situ TRDC/CD3e labeling to argue gd T cells localize outside germinal centers and are minor contributors to germinal-center processes.
Strongman: lymph node βfollicle-likeβ structure reconstruction in CSOmapR is an artifact-free proxy for true spatial organization. Why itβs less favored: reconstruction is explicitly derived from inferred ligandβreceptor expression and depends on humanized ligand/receptor gene sets, and stromal/epithelial cells were not captured, which can alter network geometry.
Quick Explanation
Long Explanation
Top Data Sources
ExportMCP
Keep Exploring
Analysis Wizard
Hypothesis Graveyard
Potential Experiments
Discussion
Explore by Goal
Ask a Follow-Up
Key Insight
