Core facility, freelancer, or analysis service: choosing who analyses your data
An honest comparison of the four ways to get single-cell data analysed — including when each one is the wrong choice.
Do this on your own data: whichever route you pick, compare the same scope — that is where the price differences turn out to live. The cost calculator itemises the scope so you can hold the comparison fixed: omics.hstgenomics.com/cost-calculator. A full deliverable to compare against what each option offers: omics.hstgenomics.com/sample.
Want it done rather than explained? We do analysis only — send a count matrix, an h5ad or Seurat object, Cell Ranger output or a public accession and get publication-ready results with the code that produced them, at a fixed price agreed before any work starts: omics.hstgenomics.com/quote. The same numbers are available to software as JSON, no key and no sign-up: omics.hstgenomics.com/api/quote.
There are four realistic options and each is genuinely the best choice in some situation. Here is how they differ on the things that decide a project: speed, cost, rigour, and what happens six months later when reviewers come back.
Your own lab
Best when the analysis is the science — a new method, an unusual assay, or work that needs tight iteration with the bench.
Worst when the analysis is standard and the person doing it will do exactly one in their career. Three to six weeks of a postdoc's time goes into learning a pipeline that will be forgotten, and the resulting analysis is usually the least reproducible of the four options because nobody wrote down the parameters.
Institutional core facility
Best when you are not in a hurry, the work is standard, and institutional billing makes life easy. Cores know your institution's data and are accountable in a way an anonymous vendor is not.
Worst when you have a deadline. A three-to-eight week queue is normal in busy periods, and no amount of budget moves you up it. Capacity is also uneven across methods — many cores are excellent at bulk RNA-seq and thinly staffed for spatial.
Freelance bioinformatician
Best when you need flexibility, a specific skill, or an ongoing part-time collaborator who learns your system.
Worst when continuity matters. One person can be unavailable exactly when the revision arrives. Rigour also varies more than anywhere else: ask directly how they would test differential expression across subjects — if the answer does not involve aggregating to subject level, keep looking.
Analysis service
Best when the work is standard, the deadline is real, and you want a fixed price and a defined deliverable rather than an hourly meter.
Worst when the project needs deep involvement in the biology, weekly discussion with the bench, or method development. A service is a good way to buy execution, not a substitute for a collaborator.
The questions that actually separate them
- How will differential expression across subjects be tested? The single best proxy for rigour. Why it matters, and what the difference looks like on real data.
- Will I get the code and the environment? If not, the next analysis starts from zero and the revision costs full price.
- Who is available in six months? Reviews come back late. Ask before you start.
- What happens if a sample fails QC? The good answer involves telling you early and adjusting the scope.
- What happens to my data afterwards? Get the deletion policy in writing, especially for unpublished human data.
A reasonable default
Many groups run a hybrid: the core or an in-house student for exploratory work where iteration matters, and a fixed-price service for the standard pipeline and for revision analyses under deadline. That combination keeps the science in the lab and the deadline risk outside it.