How to Export and Analyze QuantStudio qPCR Data

The fastest path from a QuantStudio run to usable results: export your data as a .xlsx Results file from Design & Analysis Software, pull the Ct values from the "Results" tab, and run your ΔΔCt calculations in a spreadsheet or analysis tool. That covers 90% of use cases. But the details of how you export — which settings, which file format, which columns you actually need — matter more than most people realize, because QuantStudio's export options are cluttered and the defaults aren't always what you want.

If you're working with a QuantStudio 3, 5, 6 Pro, or 7 Pro/Flex, the software is essentially the same (Thermo's Design & Analysis Software, now on version 2.x, or the older cloud-based version). The export workflow is consistent across instruments, with minor UI differences. Here's how to get your data out cleanly and what to do with it once it's in front of you.

Exporting from Design & Analysis Software

After your run finishes and the software has auto-analyzed (or you've manually adjusted the baseline and threshold), go to the Export button — it's in the upper toolbar. You'll get a dialog with several options. Here's what matters:

1. File format: Choose .xlsx unless you have a specific reason not to. The .csv export works but splits data across multiple files, which is annoying. The .eds file is the native project file — keep it as your raw archive, but don't try to analyze from it directly.

2. Content selection: Check "Results," "Amplification Data," "Melt Curve Data" (if running SYBR), and "Sample Setup." You can export everything, but the Results tab is where your Ct values live.

3. What the Results tab contains: Each row is one well. Key columns are Well Position, Sample Name, Target Name, CT (the threshold cycle), Ct Mean, Ct SD, Baseline Start, Baseline End, and flags like NOAMP, EXPFAIL, or HIGHSD. The CT column is your raw Ct for that individual replicate. Undetermined means the fluorescence never crossed threshold within the cycle limit (usually 40).

4. Threshold and baseline settings: Before you export, confirm these are set correctly. QuantStudio's auto-threshold usually lands somewhere reasonable, but it can behave oddly with low-abundance targets or noisy baselines. I prefer setting the threshold manually to the mid-log region of the amplification curve — somewhere the curves are parallel and clearly in exponential phase. For a typical SYBR Green assay (PowerUp SYBR, Luna Universal, etc.), this usually means a ΔRn threshold between 0.1 and 0.5 depending on your signal intensity. For TaqMan assays, 0.1–0.2 is common. The key is consistency: use the same threshold across all targets you're comparing.

A common mistake: exporting before checking for baseline drift. If you see amplification curves that dip below zero in early cycles or have an unusual upward slope before takeoff, adjust the baseline end cycle down (try 2–3 cycles before the earliest Ct in your plate). Auto-baseline works most of the time, but it fails quietly — the Ct values just shift by half a cycle and you never notice unless you look at the curves.

Cleaning Up the Exported Data

Once you have the .xlsx file open, the Results sheet is your primary workspace. Here's the cleanup process I use:

Filter out the flags. The NOAMP flag means no amplification was detected — treat those wells as negatives. EXPFAIL means the software thinks the amplification curve has an abnormal shape (often a sign of primer dimer or very late, nonspecific amplification). HIGHSD flags replicate groups where the Ct SD exceeds 0.5. These flags are advisory, not gospel — but they're a useful starting point.

Handle "Undetermined" wells. QuantStudio reports wells with no Ct as Undetermined rather than a number. Do not replace these with 40 (or whatever your cycle limit was) for calculation purposes unless you have a deliberate, justified reason. Treating Undetermined as Ct = 40 inflates your fold-change calculations and creates false quantitation from noise. If your NTC (no-template control) comes back as Undetermined, that's good — it means your assay is clean. If your experimental samples are Undetermined, that target probably isn't expressed in that sample, and you should report it as undetected.

Check replicate concordance. For technical replicates (duplicates or triplicates of the same sample-target combination), the SD should be < 0.5 Ct, and ideally < 0.3. If one replicate in a triplicate is off by more than 0.5 Ct from the other two, it's likely a pipetting error. You can justify dropping it if the other two agree well, but document your rule before you look at the data — don't cherry-pick post hoc.

Verify sample and target names. QuantStudio pulls these from whatever you typed into the plate setup. Typos here (extra spaces, inconsistent capitalization, "GAPDH" vs "Gapdh") will break any automated analysis downstream. Fix them now.

Running ΔΔCt Analysis

Most researchers use the comparative Ct method (Livak and Schmittgen, 2001) for relative quantification. The math is simple, but the assumptions trip people up.

The formula:

• ΔCt = Ct(target gene) − Ct(reference gene), calculated per sample
• ΔΔCt = ΔCt(treated) − ΔCt(control)
• Fold change = 2^(−ΔΔCt)

Assumption #1: Your primer efficiencies are approximately equal and close to 100%. This means both your target and reference gene amplify with efficiencies between 90% and 110% (slopes of −3.6 to −3.1 on a standard curve, respectively). If they don't, use the Pfaffl method (Pfaffl, 2001) instead, which corrects for unequal efficiencies:

Ratio = (E_target)^(ΔCt_target) / (E_ref)^(ΔCt_ref)

where ΔCt = Ct(control) − Ct(treated) for each gene, and E is the primer efficiency (e.g., 1.95 for 95% efficiency).

Assumption #2: Your reference gene is stable. If you're comparing across tissues, treatments, or time points, GAPDH is not automatically a safe choice. It shifts with hypoxia, metabolic changes, and cell density. For treatment comparisons within a cell line, GAPDH or ACTB are usually fine. For cross-tissue work, validate with at least three candidates (HPRT1, B2M, TBP, RPLP0) and use a stability algorithm like geNorm (Vandesompele et al., 2002) or NormFinder (Andersen et al., 2004) to pick the best one — or use the geometric mean of two stable references.

Assumption #3: You do your statistics on ΔCt values, not fold changes. Fold changes (2^−ΔΔCt) are ratios and are not normally distributed. Run your t-tests or ANOVA on the ΔCt values, which are log2-transformed ratios and behave much better statistically. Then convert the group means to fold change for presentation.

A worked example: You're measuring IL6 expression in treated vs. control cells, normalized to GAPDH.

Mean ΔCt control = 10.2, Mean ΔCt treated = 6.93. ΔΔCt = 6.93 − 10.2 = −3.27. Fold change = 2^3.27 ≈ 9.6-fold upregulation. Run an unpaired t-test on the ΔCt values (10.2, 10.2, 10.2 vs. 6.8, 7.3, 6.7) to get your p-value. In this case, it would be highly significant — the separation is large and the replicates are tight.

Pitfalls Specific to QuantStudio Exports

A few things bite people repeatedly with QuantStudio data:

• The "Ct Mean" column in the export is calculated from the plate setup grouping, not from your biological replicates. If you set up three wells of the same sample as replicates in the software, Ct Mean averages those technical replicates. That's fine. But your biological replicate averaging (the one that matters for statistics) has to happen in your analysis, not in the instrument software.

• Passive reference normalization (ROX). QuantStudio instruments use ROX as a passive reference dye to normalize for well-to-well fluorescence variation. The Ct values in the export are already ROX-normalized (derived from ΔRn, not raw Rn). You don't need to do anything additional, but be aware that if you accidentally ran a ROX-free master mix in a ROX-required block configuration, your Ct values could be unreliable.

• Multicomponent data for multiplex troubleshooting. If you're running multiplex TaqMan and seeing unexpected results, export the "Multicomponent Data" tab. This shows the raw fluorescence from each dye channel per cycle and lets you check for spectral bleed-through or poor probe signal separation — something the Results tab alone won't reveal.

• Cloud vs. desktop software. Thermo has pushed the Connect cloud platform, but many labs still use the desktop Design & Analysis Software (v2.4+). The export formats are slightly different — cloud exports may have different column headers or additional metadata columns. Either works; just be consistent across your experiment.

Skip the Spreadsheet Headaches

If you're doing this analysis in Excel, you already know the tedious part: pivoting wells into sample-target matrices, averaging replicates, propagating SD through the ΔΔCt calculation, and reformatting for figures. It's doable, but error-prone and slow once you have more than a couple of plates.

VoilaPCR reads QuantStudio export files directly — upload your .xlsx, map your samples and targets, and it handles replicate QC, ΔCt calculation, fold changes, and statistics automatically. It flags the same issues I described above (outlier replicates, unstable reference genes, Undetermined handling) so you can focus on interpreting biology instead of debugging formulas.