What does single-cell RNA-seq data analysis actually cost?

Sequencing prices are published everywhere. Analysis prices are not, which makes budgeting a grant painful. Here is what the four realistic options cost, and what each one quietly leaves out.

Doing it in-house

The apparent cost is zero, which is why it is chosen most often. The real cost is a postdoc's time: a first standard analysis takes a competent but inexperienced person three to six weeks including the learning curve, and a further one to two weeks each time reviewers ask for something. At a typical fully-loaded postdoc cost, that is several thousand dollars of salary spent on a task that is not their research contribution.

In-house wins when the analysis is the science — novel method development, unusual assay, tight iteration with the bench. It loses when the analysis is standard and the person doing it will never do another one.

Institutional core facility

Typically charged as an hourly rate, commonly in the range of 60–150 USD per hour, or as a per-project package. A standard scRNA-seq project usually lands between 1,000 and 4,000 USD. Quality is generally good and the people know your institution's data.

The constraint is the queue. Cores serve everyone, and a three-to-eight week wait is normal in busy periods. If your revision deadline is in four weeks, the core may not be able to help even if the price is right.

Freelance bioinformaticians

Rates of 40–120 USD per hour are common, so a standard project lands roughly between 800 and 3,000 USD. Fast and flexible. The risks are concentration risk (one person, who may be unavailable when review comes back) and variance in rigour — pseudobulk versus per-cell differential expression, for instance, is exactly the kind of methodological choice that separates a defensible analysis from one that collapses under review.

Analysis service providers

Fixed-price packages, usually 300–700 USD per sample for a standard single-cell analysis, with advanced modules (integration, trajectory, cell–cell communication, differential abundance) adding roughly 900–3,000 USD at the project level. Spatial slides typically run 550–1,400 USD each because deconvolution and niche analysis are heavier. Bulk RNA-seq differential expression projects are far cheaper, commonly 300–900 USD.

The reason to use one is turnaround and predictability, not price. The reason to be careful is that "analysis" means very different things to different vendors.

The hidden costs, in order of how often they bite

1. Revisions. Ask before you sign: are reviewer-driven re-runs included, and for how long? A package that excludes revisions is not a package, it is a first draft.
2. The processed object. If you only receive figures and tables, the next analysis starts from zero. Insist on the h5ad or Seurat object plus the script.
3. Figure formats. Journals want vector or 300 dpi. Screenshots of a notebook are not a deliverable.
4. Methods text. Someone has to write the paragraph with tool versions and parameters. If it is not in the quote, it is on you, at the worst possible moment.
5. Data transfer. Moving 200 GB of FASTQ across the world is slow. Working from count matrices, where possible, removes days.

A rule of thumb for budgeting

For a standard project of 4–8 samples, budget 8–12% of what you spent on the sequencing itself for the analysis. Below that, you are either doing it yourself or buying something narrower than you think. Well above it, you should be getting method development, not a pipeline run.