Field overview
Single-cell biology
Bulk assays — RNA-seq, ATAC-seq, run on millions of cells at once — measure an average that can hide everything interesting. A 50/50 mix of two cell states looks identical to a uniform intermediate state in bulk.
Single-cell methods resolve that. You get a measurement per cell, which lets you find rare populations, reconstruct developmental trajectories, and map heterogeneity in tumors or tissues.
The big threads
04- 01
scRNA-seq
The workhorse. Droplet platforms tag each cell’s transcripts with a cell barcode and UMI. You end up with a cells × genes count matrix that’s huge, sparse, and noisy.
- 10x Genomics
- dropout
- ambient RNA
- doublets
- 02
Multi-omics
Multimodal readouts — RNA plus chromatin accessibility, or RNA plus surface protein — measured in the same cell.
- scATAC
- Multiome
- CITE-seq
- 03
Spatial transcriptomics
Keeps the location of cells in tissue, the information that single-cell dissociation throws away.
- Visium
- Xenium
- MERFISH
- 04
The computational side
Where the day-to-day work lives: QC, normalization, dimensionality reduction, clustering, batch integration, trajectory inference, differential expression, and cell-type annotation.
- scanpy
- Seurat
- UMAP
- Leiden
- Harmony
- scVI
Where to take this
Four directions, depending on what you’re after.
Conceptual
The biology and what these experiments reveal.
Analysis and computational
Working with the data, pipelines, a specific tool or method.
A specific problem
You have a dataset or a project in mind.
The state of the art
Foundation models for cells, recent methods, where the field is heading.
- scGPT
- Geneformer
What’s the context — are you learning, building something, or analyzing real data?