Keyboard shortcuts

Press or to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

Glossary

Adjusted Rand index (ARI): Agreement between two partitions, corrected for chance and independent of numeric cluster names.

Ambient RNA: RNA released into the suspension and captured by droplets that may not contain the source cell.

Barcode: Sequence used to associate molecules with a captured droplet or cell library.

Batch: A group sharing a technical processing event. Batch can be confounded with biology.

Cell state: A relatively temporary program such as activation, stress, or cell cycle.

Cell type: A biological identity supported by lineage, function, markers, and context. It is not identical to an algorithmic cluster.

Cluster: A group produced by an algorithm under selected data and parameters.

Count matrix: Table of observed molecule counts for cells and genes.

Differential expression: A statistical comparison of expression between defined groups under a specified design.

Doublet: One barcode representing two captured cells.

Experimental unit: The smallest independently assigned or sampled unit that supports inference, often a donor, animal, organoid, or culture.

Feature: A measured item, usually a gene but potentially an antibody tag or guide.

Gene: A genomic region that contributes to a functional RNA or protein product.

Highly variable gene (HVG): Gene selected because its variation is useful for describing cell structure under a particular method.

Integration: Construction of a representation intended to align shared biology across samples or batches.

Library size: Total observed counts for one cell library.

Marker: Gene whose expression helps distinguish a group. A marker is context-dependent and not necessarily unique.

Mitochondrial fraction: Fraction of a cell’s counts assigned to mitochondrial genes.

Neighbor graph: Graph connecting each cell to similar cells in a selected representation.

Normalization: Transformation intended to make measurements more comparable for a downstream purpose.

PCA: Principal component analysis, a linear low-dimensional representation.

Pseudobulk: Counts aggregated across cells within each biological sample and cell type for replicate-aware analysis.

Pseudoreplication: Treating correlated observations, such as cells from one donor, as independent experimental replicates.

Pseudotime: Relative ordering along an inferred cellular continuum. It is not clock time.

RNA: Molecules involved in gene expression and regulation. Messenger RNA is used as a proxy for transcriptional activity.

Sparse matrix: Matrix stored using nonzero entries rather than every zero.

Transcript: RNA molecule produced from a gene.

UMAP: Nonlinear low-dimensional visualization. Its visual distances are not calibrated biological distances.

UMI: Unique molecular identifier used to reduce duplicate counting of the same captured molecule.