Understanding Malva¶
This section introduces the ideas behind Malva before you start using it. It explains how the data is organised, what a search does, and how the results are put together, so that the rest of the documentation reads naturally.
What Malva is¶
Malva is a search engine for single-cell and spatial transcriptomics data. It indexes published single-cell studies, and you can search across them by the sequences expressed in their cells. You can also browse and filter the data by its metadata, compare cell groups, and share what you find.
Malva comes in three parts:
- The web application, where you search, browse, and analyse in the browser.
- The HTTP API, which exposes the same operations to your own programs.
- The Python client, a package named
malva-clientthat wraps the API.
How the data is organised¶
The database contains single-cell samples from published studies. The terms below appear throughout the documentation and in the interface.
| Term | Meaning |
|---|---|
| Sample | One biological specimen measured in a study. A sample can contain thousands of cells. |
| Dataset | A coherent collection of samples loaded into Malva, usually a single atlas. |
| Study | A published study that contributed one or more samples. A study belongs to a dataset. |
| Cell type | The annotated cell type of a cell, such as T cell or hepatocyte. |
| Search | A single query submission. Each search consumes one unit of your daily search quota. |
| Job | A container that holds one or more searches and their results. Jobs appear in your job history. |
| Artifact | One sequence within a job. A single search can carry up to 10 sequences. |
| Probe | A 48-base window that the Coverage Explorer slides across a region to measure coverage. |
| Metacell | A compact summary of many similar cells, used to keep group analysis fast. |
Every cell in the database has an expression profile, and every sample carries metadata: its organ, disease associations, sex, species, study, and more. Malva uses that metadata for filtering and for ontology-aware browsing, described in Ontologies.
The three views¶
The web application provides three working areas:
- The Expression Explorer searches for genes and sequences and shows which cell types and samples express them.
- The Coverage Explorer measures read coverage across a genomic region or a pasted DNA sequence.
- The Analysis workspace saves cells as groups and compares them with marker-gene and differential-expression analyses.
The tutorials for each view build on the concepts in this section.