2026.10·Run it your way

All releases · October 12, 2026 · Jan Hellemans

Last month was about getting your data in and making the calls yourself. This month is about running every experiment the way your lab works, without repeating the setup each time: the sections an experiment needs, the QC thresholds you trust, the settings a kit expects.

That setup now lives in the Library. A workflow layout decides which sections an experiment shows and in what order. A settings set decides the thresholds and choices behind the numbers. A preset carries both, so starting an experiment from one gives you a workflow that already fits the question you are asking.

The first workflow to use this from day one is a new one: genotyping by allelic discrimination. And if you have ever lost your place in a long workflow, the new Navigator shows where you are and takes you anywhere in it.

Find your way.
Jan

New features

Genotyping goes beyond Cq values

Since the previous release Clarida has handled your raw data, but every analysis still started from Cq values. Genotyping changes that: besides ΔCq, it uses melting temperatures (Tm) and end-point fluorescence to call alleles, with a published calling method. You declare each marker once (its alleles, which target and dye reads which allele, and the method), and Clarida turns the measurements into a sample-by-marker genotype table.

The Genotype calling section for marker rs16911905, an end-point marker with two alleles, Mutant in HEX and WT in FAM. The Settings card calls in X–Y space with cut-offs of 600 RFU for Mutant and 780 RFU for WT, and the Quality control card reads Attention, with 7 clean calls and incomplete controls. Below, a cluster plot of WT against Mutant fluorescence shows three groups separated by the two cut-offs (WT/WT, Mutant/WT and Mutant/Mutant), beside a table listing each sample's Mutant and WT signal, θ, R and its genotype call

End-point markers get a cluster plot, with calling in polar space. Every call comes with a quality value, and a marker Clarida cannot call with confidence is reported as declined rather than guessed. Known-genotype controls are checked against their expected calls, so a mirrored plate or swapped dye shows up as a control failure instead of a table of inverted genotypes. Markers live in the Library next to your assays, and a kit can ship its markers and expected control genotypes with it.

The full method, from declaring a marker to judging a call, is written up in Genotyping by allelic discrimination in our methods library.

Choose the workflow that fits the question

Not every experiment needs every section. A detection run has no use for normalization, and a homogeneity check needs only a fraction of the qPCR workflow. A workflow layout decides which sections an experiment shows, in what order, and how they are grouped, and you choose it when you create the experiment.

An experiment on the genotyping workflow layout. The Import and review group is open and holds only the sections genotyping needs: Import and annotate, Sample prep, Assay prep, Marker setup and Run Layout, followed by Analyze and Report. The Navigator on the right lists the same layout, with Exclusion, Genotype calling, Sample/Assay QC and Genotyping under Analyze, and Report builder under Report

Clarida ships layouts for relative quantification, detection, genotyping, reference gene selection and plate homogeneity, plus one that shows every section. You can also build your own in the layout editor: leave out what you never use, move analysis sections where they make sense to you, and gather sections in named groups.

Finding your way through a long workflow got easier too. Closed sections take less room, and the breadcrumb at the top extends to the section you are reading. A new Navigator tab, now the first tab of the side panel, shows the whole workflow as a tree with each section's status, and a click takes you there. Press Ctrl+K and type a section's name to jump to it from anywhere.

Define your settings once

QC thresholds, efficiency handling, reference-gene criteria: every lab has its own, and until now every experiment started from Clarida's defaults until you changed them. A settings set is a named, versioned bundle of those choices that lives in the Library. Every experiment started from a preset that carries it begins with exactly those settings.

Duplicating a settings set. It starts from Clarida defaults v1 and is named 'Clarida defaults for 384-well'. With 'Only differing from parent' switched on, the settings table shows the single change: Plate format, in the Design group for all applications, set to 384-well where the parent value is 96-well, with a reset action beside it. The header reads '1 setting changed' next to Cancel and Create

Each experiment records the set it started from, and Define and the ELN export name it, so a reviewer can see which configuration produced the numbers. Clarida's own defaults are now a set like any other, called Clarida defaults, which you can inspect setting by setting. A lab set derived from it shows exactly where it differs, and when the parent gets a new version you can review the changes and adopt only the ones you want.

Start from your favorite presets

The qPCR landing page now opens with a shelf of one-click presets: the ones you pinned, or else the ones your workspace chose, or else your starred presets topped up with a selection of Clarida starters. Since a preset now carries its workflow layout and settings set, one click gives you an experiment that is configured, not just created.

The 'Start a qPCR experiment' page. Below the two entry cards, Build from scratch and Start from data you already have, a 'Start from a preset' shelf offers eight one-click Clarida presets: Relative quantification, Relative quantification – analysis only, Reference gene finder, Reference gene finder – analysis only, Plate homogeneity check, Genotyping, Genotyping – analysis only and Detection, with Customize and All presets links. The Continue list of experiments follows below

Improvements

Share an instrument connector between workspaces · A Porta connector can now serve more than one workspace. The workspace that owns it stays in control and decides which other workspaces may use it, so one instrument PC can serve several groups without mixing their data.
Porta on the instrument PC · The connector's setup page shows which workspace and connector it belongs to, and the machine's owner can disconnect it locally. Workspace admins get an email when that happens, and every change email now names the connector involved.
See where a copy came from · Samples, assays, mixes and markers copied from the Library or a kit show their origin. Once you edit one, it becomes the experiment's own copy: the origin gives way to Add to library, and a reset action shows what you changed and restores the Library version.
Find Library items by source · The long vendor dropdown on the Library pages is now three filters: Mine, Clarida, and Vendors.
A tidier Library · Every Library page uses the same filters in the same order, Markers get a proper list view, and the sidebar groups Presets, Workflow layouts, Experiment settings and Connectors under Setup. The Experiments list shows which preset each experiment started from.
Sortable, filterable tables everywhere · The column headers introduced in 2026.9 (sort, filter, resize, freeze, reorder) now reach the remaining tables in Statistics, Exploration, Cq processing, Import, Raw data QC, QC and Detection. Number columns sort as numbers.
Melt curve or melt peak, smoothed your way · The Melt view now offers the two views that exist, the melt curve and the melt peak (−dF/dT), with smoothing as an option on the peak view instead of a third view. The smoothing window is Auto by default or set by hand, and a warning names the wells where it falls below the minimum.
A faster app · Opening an experiment, importing a 384-well plate and browsing the Library are all markedly faster. Some steps went from seconds to a tenth of a second.
ELN export in seconds · A typical ELN export takes under 2 seconds instead of about 25, with identical output, and no longer slows down the app for others while it runs.

Fixes

Bio-Rad CFX RDML files import · RDML files exported from a Bio-Rad CFX were refused. They now import like any other RDML file.
RDML exports show real names · Exported RDML files used internal database IDs as sample and target names, so qbase+ and RDML-Ninja showed codes instead of names, and re-importing created duplicates. Names are now exported as RDML IDs.
Long run names no longer stop an import · A run file whose name exceeded the stored length failed at the last import step. Long names are now accepted.
Unused wells no longer block Reconcile · Empty plate positions showed up as conflicts that could not be resolved. They are now left out.
Raw data stays in step · Renaming a sample or assay now updates the raw data view, large plates no longer lose their curves on import, and OpenArray .eds files are refused instead of being read with wrong well positions.
End-point values appear whatever the layout · End-point values were only calculated while the Raw data QC section was open, so a layout without it left the end-point columns in Amplification empty. Switching a readout to imported and back to the same calculated method also left its Cq, Tm or end-point column empty. Both now calculate as expected.
"Add to library" works on manual entry · The checkbox in Sample prep and Assay prep's manual method had no effect. Items are now added to your Library.
Mix planning behaves · The mix plan recalculates when you change a scale fraction or reload, the expanded recipe is shown in full, and a stale mastermix warning no longer lingers. Fill plate showed 0.0 µL for "Add sample" on most experiments and now shows the real volume.
Run Layout lock follows the run · The lock on Run Layout now depends on whether the run has moved past planning, says why it applies, and can be released from the status badge.
Duplicated experiments are complete · A duplicate now carries everything the original holds, including saved analyses, excluded reference genes, efficiency overrides and detection cutoffs, and gets its own copies of protocols and mixes. Duplicates made before this release did not get these; check them if you rely on one.
Experiments open in the right workspace · Opening a link to an experiment from another of your workspaces now switches to that workspace and says so.
No more false "Experiment not found" · Returning to a tab that had been idle could replace a loaded experiment with a permanent "Experiment not found". Clarida now reconnects, or asks you to sign in again.
Sign-up completes · Creating an account could time out while the demo data was being prepared. Demo data is now added in the background, and a failure during onboarding shows a way forward.
Correlation trendlines on a log axis · On a log axis the trendline could be drawn far from the data. It now follows the scale the fit was made in. The fitted values were always correct.
Large experiments export · The report failed on large plates and the ELN export could return an error on large experiments. Both now complete.
Readable charts in dark mode · Axis labels, brackets and reference lines were nearly invisible on dark backgrounds, including the NRQ chart.
Clearer billing messages · When adding or removing a card fails, the message says why, instead of suggesting that only the owner may do it.

Changes that may affect you

The calculations themselves did not change: Cq calling, normalization and statistics work exactly as before. What changed is which inputs reach them, and how one verdict reads them, in four situations, three of which were bugs. None of these changes recalculates an existing experiment on its own: the numbers change the next time the experiment is recalculated, and mix volumes the next time it is opened.

A library efficiency was shown but not used. For a target with an amplification efficiency in your Library and no standard curve in the experiment, Clarida displayed that efficiency but calculated relative quantities with an efficiency of 2. The calculation now uses the efficiency it displays, following the priority you chose in your settings. If you use library efficiencies, your NRQ values for those targets will change at the next recalculation.

Mixes in experiments with several pipetting protocols used the default protocol. Every mix group got the recipe of the experiment's default protocol instead of its own. Each group now follows its own protocol, and the ELN and Excel exports pair each protocol's overage with its own volumes. Affected experiments show the corrected volumes the next time you open them.

The reference-gene CV indicator has no amber band anymore. The stability indicator for reference genes was green up to 25%, amber up to 50% and red above. It now uses one cutoff, taken from your settings (25% by default): green or red. A reference gene between 25% and 50% now shows red.

Reference-gene stability now matches qbase+. The geNorm M values in Normalization were calculated as if every target amplified with an efficiency of 2. They now use the efficiency-corrected quantities, as qbase+ does, so for targets with an efficiency other than 2 the M values change slightly at the next recalculation. The "not stable enough" warning changed too. It fired as soon as a single reference gene's M exceeded the cutoff, and it ignored CV. It now judges the mean M and mean CV of the included reference genes, as described by Hellemans et al. (2007). The values per gene are still shown. A verdict close to the cutoff may change.

Separately, a fix to Measurement exclusion: after changing a filter, "Exclude selected" could exclude a measurement you had not selected. That no longer happens, but exclusions made that way before this release remain. If you used filters together with "Exclude selected", check your exclusions.