MSCinteract documentation

MSCinteract is an interactive explorer for mesenchymal stromal cell (MSC) expression datasets and the indications they are studied in. Every matrix is streamed from the server on demand, parsed and analysed inside your browser — nothing is uploaded, and results appear immediately.

What MSCinteract is

MSCinteract is an interactive explorer for the expression datasets of mesenchymal stromal cells (MSCs) and the indications they are studied in. It is organised as a few tabs: the QuickCheck interfaces let you question one matrix at a time — the 20-media culture of UC-MSCs, or one of the five disease cohorts — and the Interaction interface shows how the MSCs and the immune cells talk to each other. The catalogue on the home page gathers everything that is published, and more is on its way.

• You can ask for one gene or a hundred — type a symbol, paste a whole list, or load a curated panel of MSC biomarkers, immunomodulatory mediators and receptors.
• You can read every answer as a picture or as a number — bar chart, heatmap and sample-correlation scatter on the picture side, the genes × samples table and exact per-sample values on the number side.
• You can move between conditions and cohorts without writing code — the same explorer runs on the culture conditions, on COVID-19 lung, GvHD, IBD, IPF and KOA, and correlates any two samples gene by gene.
• You can follow the conversation — in the Interaction tab, the dot plot tells you which genes carry the signature of a condition, and the bubble plot which immune cell types can answer.

Quick start

The explorer runs directly in the browser: matrices are streamed from the server on demand and analysed on the fly. Three flows cover everything the site does — find a gene in the MSCs, find a gene in a disease cohort, and follow the talk between the MSCs and the immune cells.

A gene in the MSCs

A gene in a disease cohort

MSC–immune cell talk

QuickCheck interface

QuickCheck is the expression explorer of the site. One matrix of genes × samples is open at a time — the 20-media culture of UC-MSCs in the MSCs tab, one of the five reference cohorts in the Diseases tab — and everything you do is a question to that matrix: which samples express this gene, how do two samples compare, which genes switch on together.

Open a matrix

The MSCs tab opens the 20 culture conditions of UC-MSCs. In the Diseases tab, the picker in the header card switches to COVID-19 lung, GvHD, IBD, IPF or KOA, and the matrix downloads in the background while you keep working. The chips in the header card tell you at any moment how large the open matrix is — how many genes, how many samples — and Reload fetches a fresh copy whenever you want one. Switching cohorts resets the gene selection, because every cohort has a gene index of its own.

Bring your genes in

Type a symbol such as TP53 into the search box of the sidebar: the dropdown suggests matches as you type and Enter takes the first hit. To bring a whole list, paste symbols separated by commas, spaces or line breaks and press Import list — symbols the matrix does not know are reported back to you instead of vanishing. Every selected gene appears as a removable chip, and the counter above tells you how many the selection carries. The ready-made panels — MSC biomarkers, Immunomodulatory and Receptors — load in one click and mix freely with your own list.

Reading the charts

Click the row of a gene in any chart or table to make it the focus: the statistics above the plots then show its mean, median, maximum, minimum and the number of samples that express it.

Bar chart — one gene across all samples

Each bar is one sample of the dataset and its height is the expression value of the focus gene. Use Bar order to switch between the natural sample order (which usually follows the experimental grouping) and a ranking by value.

Heatmap — genes × samples

The heatmap shows the selected genes (up to 60 at a time) across all samples. Rows are genes, columns are samples, and the colour scale is relative to the values of the current selection, which makes patterns — for instance a group of samples that switches a pathway on or off — easy to spot.

Scatter — correlation between two samples

The scatter plot compares two samples gene by gene and reports the Pearson correlation. By default all genes of the matrix are used; switch the scope to selected genes to correlate only the genes in your current selection. Because the correlation needs the complete matrix, the missing shards are fetched on demand when you open this chart.

Reading the tables

Per-sample expression lists every sample of the dataset with the value of the focused gene. Use it to read exact numbers for the samples that matter, for example a treated versus untreated pair.

Expression matrix shows the selection as a genes × samples table. Click a column header to sort by that gene, click a gene row to make it the focus gene, and use the pager to move through long selections.

Interaction interface

This tab asks how MSCs and immune cells talk to each other. The MSC side is the 20-media dataset of QuickCheck: MSCs; the immune side is the matrix of 15 immune cell types from the DICE database; and the pairs of ligands and receptors that connect the two sides come from OmniPath. The tab does not compute anything on a server — it reads the marker panels of the selected condition, streams the matching expression values and compares them in your browser.

Pick the direction and the condition

MSCs → Immune cells follows the factors the MSCs secrete to the receptors they act on in the immune cells; Immune cells → MSCs starts from the receptors of the MSCs and follows the immune ligands that reach them. The panel titles and the hint under them follow your choice. In Query mode you pick the one culture condition the interaction is built on; in Comparison mode you pick a second one, and the tab tells you gene by gene what the two conditions share and what belongs to only one of them.

Read the two plots

The dot plot draws every marker gene of the selected condition across all 20 media conditions (log2(TPM + 1)), with the selected condition highlighted; the chips in the legend mark the genes that top the condition and the ones unique to it — in Comparison mode, the ones the two panels share and the ones only one of them carries. Hover the name of a gene to see the partners it works with on the other side. The bubble plot puts the same genes against the 15 immune cell types: the bigger and darker the bubble, the stronger that cell type expresses the partners — that is, the more likely it hears the MSCs. In Comparison mode the rows are grouped into common, only in the first and only in the second condition, with one colour per group. Read the dot plot first — it tells which genes carry the signature — then the bubble plot, which tells who can answer.

Units and values

• TPM — transcripts per million. The unit of every cohort on the site, including the MSC media dataset; comparable between samples of the same dataset.

Data sources

Every matrix of the site is listed below with its content, its size, where the data lives and every publication to cite when you use it. The MSC media dataset runs in the QuickCheck: MSCs tab and the five cohorts in the QuickCheck: Diseases tab; the last two sources serve the Interaction tab.

UC-MSCs Cultured in Different Media

• The dataset — 20 media conditions (our defined media, FBS, NBCS, BS, DBS and hPL); 20 samples × 43,737 genes, TPM.
• Origin — produced in house (University of Macau / MacOrigen Macau). The dataset has not been deposited in a public archive and no paper accompanies it yet; the accession and the citation will be added here as soon as it is published.

Sample names

COVID-19 Lung Lobes — GEO GSE150316

• The dataset — sampling sites (lung lobes LLL/LUL/RLL/RML/RUL) of COVID-19 patients, each site split by the viral load of the sampled tissue (High/Low), plus healthy controls; 47 samples × 54,357 genes, TPM.
• Desai N, et al. Temporal and spatial heterogeneity of host response to SARS-CoV-2 pulmonary infection. Nature Communications 2020;11:6319. doi:10.1038/s41467-020-20139-7 · PMID 33298930

Sample names

GvHD Lymphocytes — GEO GSE182610

• The dataset — peripheral blood lymphocytes (T cells), by disease status and subset; 37 samples × 55,845 genes, TPM.
• McCurdy SR, et al. Signatures of GVHD and relapse after posttransplant cyclophosphamide revealed by immune profiling and machine learning. Blood 2022;139(4):608–623. doi:10.1182/blood.2021013054 · PMID 34657151

Sample names

IBD Colon and Ileum Explants — IBDome

• The dataset — inflammatory bowel disease tissue cohort, harmonised to gene symbols; 414 samples × 55,512 genes, TPM.
• Official site — ibdome.org, the explorer of the TRR241 research initiative; the underlying biopsies are not served as a single GEO series.
• Plattner C, et al. IBDome: an integrated molecular, histopathological, and clinical atlas of inflammatory bowel diseases. Gastroenterology 2026;171(5):902–918. PMID 42269946
• Plattner C, et al. IBDome: an integrated molecular, histopathological, and clinical atlas of inflammatory bowel diseases. bioRxiv 2025:2025.03.26.645544. doi:10.1101/2025.03.26.645544 · PMID 40291692

Sample names

IPF and CHP Lung Tissue Biopsy — GEO GSE150910

• The dataset — idiopathic pulmonary fibrosis biopsies (Bx) and explants (Ex), plus healthy controls and the chronic hypersensitivity pneumonitis (CHP) cohort; 288 samples × 17,819 genes, TPM.
• Furusawa H, et al. Chronic hypersensitivity pneumonitis, an interstitial lung disease with distinct molecular signatures. American Journal of Respiratory and Critical Care Medicine 2020;202(10):1430–1444. doi:10.1164/rccm.202001-0134OC · PMID 32602730

Sample names

KOA Knee Synovium — GEO GSE283079

• The dataset — knee osteoarthritis synovium (36 KOA) and healthy controls (5); 41 samples × 59,453 genes, TPM.
• Miyahara J, et al. CD34^hi subset of synovial fibroblasts contributes to fibrotic phenotype of human knee osteoarthritis. JCI Insight 2025;10(2):e183690. doi:10.1172/jci.insight.183690 · PMID 39846253

Sample names

Immune cells of the Interaction tab — DICE

• Official site — dice-database.org, the database of immune cell expression, eQTLs and epigenomics.
• Accession — the 15 sorted immune cell types of the DICE bulk RNA-seq matrix.
• Schmiedel BJ, Singh D, Madrigal A, et al. Impact of genetic polymorphisms on human immune cell gene expression. Cell 2018;175(6):1701–1715.e16. doi:10.1016/j.cell.2018.10.022 · PMID 30449622
• Chandra V, Bhattacharyya S, Schmiedel BJ, et al. Promoter-interacting expression quantitative trait loci are enriched for functional genetic variants. Nature Genetics 2021;53(1):110–119. doi:10.1038/s41588-020-00745-3 · PMID 33349701
• Schmiedel BJ, Rocha J, Gonzalez-Colin C, et al. COVID-19 genetic risk variants are associated with expression of multiple genes in diverse immune cell types. Nature Communications 2021;12(1):6760. doi:10.1038/s41467-021-26888-3 · PMID 34799557
• Schmiedel BJ, Gonzalez-Colin C, Fajardo V, et al. Single-cell eQTL analysis of activated T cell subsets reveals activation and cell type-dependent effects of disease-risk variants. Science Immunology 2022;7(68):eabm2508. doi:10.1126/sciimmunol.abm2508 · PMID 35213211

Ligand–receptor pairs of the Interaction tab — OmniPath

• Official site — omnipathdb.org, the literature-curated resource of signalling pathways; the pairs of the tab are its ligand–receptor table.
• Türei D, Korcsmáros T, Saez-Rodriguez J. OmniPath: guidelines and gateway for literature-curated signaling pathway resources. Nature Methods 2016;13(12):966–967. doi:10.1038/nmeth.4077 · PMID 27898060
• Türei D, Valdeolivas A, Gul L, et al. Integrated intra- and intercellular signaling knowledge for multicellular omics analysis. Molecular Systems Biology 2021;17(3):e9923. doi:10.15252/msb.20209923

Questions and corrections

Questions, corrections or requests for additional cohorts can be sent to data@macorigen.com.

FAQ