qbase+ was retired after CellCarta acquired Biogazelle in 2021; by 2024 the software was no longer maintained, supported or downloadable. That much is documented by the people who built it. What is less documented is what to do next if your methods section says “normalised using qbase+”, your raw files are still on a lab drive, and the reviewer wants to know the numbers are still what you said they were.
The person who wrote the software put the problem better than any vendor page does:
“The question wasn't just ‘what tool do I use?’. It was ‘how do I keep doing this correctly?’”
Jan Hellemans, co-founder of Biogazelle and of Clarida
That distinction matters, because the four paths below answer different questions. Only one of them is about “a tool”. The rest are about what you are allowed to do with your data.
1. Clarida — the same team, rebuilt around the same methods
Clarida is the direct successor: geNorm for reference-gene selection, efficiency-corrected normalisation, inter-run calibration, MIQE-oriented reporting, built by the team that wrote the originals. As of its 2026.6 release it states that the analyses qbase+ had are covered, including inter-run calibration and sample QC, and it adds assay QC and global-mean normalisation for large panels. Its free tier includes a geNorm tool, an RDML viewer and an RDES validator.
Choose this if: you want the whole pipeline — design through report — and uploading your files to an EU-hosted service is acceptable to your institution. It is genuinely well built, and it is the least-effort path back to a normalisation workflow that resembles the one you lost.
Do not choose it if your data cannot leave the machine (unpublished, patient-derived, pre-patent, or under an institutional rule that says so), or if per-experiment billing would make you hesitate every time you analyse something. Clarida's pricing was listed at €15 per experiment when we checked in October 2026; verify it yourself before you budget.
2. VoilaPCR — cheapest path to fold changes and figures
A browser tool aimed at the workflow most people actually run: import the instrument export, get ΔΔCt, statistics, publication figures. It supports ΔΔCt, Pfaffl, geNorm and standard curve methods, and exports figures at 300 DPI plus Prism files. When we checked in October 2026 the page showed a $99 list price struck through to $20/year after five free analyses.
Choose this if: your need is “get from a plate export to a figure and a statistics table, cheaply, today”, and you do not need to defend how each Cq was derived.
One thing to verify yourself: VoilaPCR states that processing happens in your browser and the data never leaves your computer. That is a claim you cannot audit from outside. If it is the deciding factor for you, ask the vendor for the network trace, or run it once with the network disconnected and watch what happens.
3. RDML plus free tools — the zero-cost, high-effort route
If your instrument exports RDML, or you can export it, the free ecosystem is real and good: rdml-python for parsing, RDML-Ninja for inspection and repair, qPCRDataMethods for a benchmark across nine published Cq methods, LinRegPCR for efficiency-corrected Cq, and a geNorm implementation in R.
Choose this if: you have someone who will actually maintain a script, and your institution has no budget line for software. The cost is not money, it is the week you spend getting the export shape right — see the trap below, which is where most of that week goes.
4. An offline desk tool — what this site is building
CqVerify reads the vendor files directly (.pcrd, .eds, .ixo, .rex), recomputes Cq by three independent methods, prints every parameter it used, and produces a report built around the MIQE 2.0 reporting fields. It does not connect to anything. There is no account, and nothing is uploaded, as an architectural property rather than a policy.
It is in development, with no promised release date, and it is the right answer for a narrow case: you need to show how a number was produced, and you are not allowed to upload the file to find out. If neither of those is your situation, options 1–3 will serve you better and cost you less time.
| Your constraint | Pick | Why |
|---|---|---|
| Cloud is fine, want the full pipeline | Clarida | Same authors, same methods, widest coverage |
| Need a figure and statistics cheaply | VoilaPCR | Lowest price, shortest path to output |
| No budget, have a programmer | RDML + free tools | Free, reproducible, fully under your control |
| File cannot leave the machine | CqVerify (in development) | Offline by architecture, not by promise |
| Reviewer asks how the Cq was derived | CqVerify or a printed parameter log | Requires per-run parameter disclosure |
The trap that costs everyone a week: export shape
Whatever path you take, you will meet the same wall. Most downstream tools assume the data they receive is raw fluorescence. Most instrument exports default to baseline-corrected data. The author of LinRegPCR says it plainly in the official training material:
“Keep in mind that also the RDML export of the qPCR machine should contain raw, not baseline-corrected, data.”
LinRegPCR training material
Note the word “should”. When it does not, users meet errors like
list index out of bounds or
cannot convert variant of type (Olestr) into type (Double), and conclude
that the tool is broken. The tool is fine; the export was the wrong shape.
This is checkable before you analyse anything, and it is worth ten minutes at the start of a migration: open the export, look for negative values, ask whether baseline correction has already been applied, and re-export as raw if your method needs it.
What “correctly” means in 2026, if a reviewer is involved
Reporting Cq and efficiency is no longer enough to let a reader judge bias. From Bustin 2023: reporting quantification cycle and PCR efficiency values is insufficient to enable reviewers or readers of a paper to assess bias. And from MIQE 2.0, seventeen years after MIQE 2009 raised the same point: “MIQE is routinely ignored.”
So the practical standard for a replacement tool is not “does it produce a number”. It is “can you print the parameters that produced the number, and can someone else reproduce it from what you printed”. Any of the four paths above can meet that standard. Only you know which one your institution lets you use.
Keep one plate from the old qbase+ era, run it through whichever tool you choose, and compare the normalised quantities rather than the Cq values. Inter-run calibration differences show up in the normalised result, not in a single well. If those match, the migration is done. If they do not, the cause is almost always calibration settings, not the algorithm.
Prices and product claims were checked on 20 October 2026 against each vendor's own pages and release notes, and are stated for comparison, not endorsement. Clarida and qbase+ are products of their respective owners; MIQE and geNorm refer to published works. If a figure here is out of date, write to [email protected] and it will be corrected rather than quietly deleted.