ChemiDoc vs LI-COR Odyssey for Western Blot Quantification

If you're choosing between a Bio-Rad ChemiDoc and a LI-COR Odyssey for quantitative westerns, the single biggest difference is detection chemistry, not the box itself. The ChemiDoc captures chemiluminescent signal (ECL) that decays over time and is generated enzymatically — meaning the relationship between protein amount and signal is nonlinear once substrate is depleted or HRP is saturated. The Odyssey uses direct near-infrared (NIR) fluorescence from secondary antibodies conjugated to IRDye 680 or 800, which gives you a stable, non-enzymatic signal with a wider linear dynamic range. For quantification specifically, the Odyssey wins on paper. In practice, either system can produce defensible data if you understand its limitations.

Here's the short version: if your lab already has a ChemiDoc and you're doing straightforward fold-change comparisons with 2–3× differences, you can get solid quantification with careful exposure management. If you're setting up a new lab, doing multiplex detection, or chasing small (< 2×) changes, the Odyssey's ~4-log linear dynamic range and two-color capability will save you headaches. Now let's unpack the details.

Dynamic Range: Why It's the Only Spec That Really Matters

The practical linear dynamic range of chemiluminescent detection on a CCD-based imager like the ChemiDoc is roughly 2–2.5 orders of magnitude under good conditions (Gassmann et al., 2009). That's a ~100–300-fold range between the lowest detectable signal and saturation. Sounds generous until you realize that a loading series from 1 µg to 40 µg of HeLa lysate — a common scenario — can easily span that range for an abundant target. And that's before you account for the fact that ECL signal is time-dependent: the enzymatic reaction follows a rise-peak-decay curve, so a band captured at 30 seconds post-substrate addition doesn't have the same intensity-to-protein relationship as one captured at 5 minutes. Bio-Rad's "signal accumulation mode" (multiple short exposures computationally stacked) in Image Lab helps, but it's correcting for a fundamental issue rather than eliminating it.

The LI-COR Odyssey CLx (or the newer M series) uses direct fluorescence excitation at 685 nm and 785 nm. No enzyme, no substrate, no decay. The signal is proportional to the number of fluorophore molecules bound, full stop. Published linear dynamic range is 4 orders of magnitude (10,000-fold), and in my hands it's reliably 3–3.5 logs with typical membrane autofluorescence. That means you can have a faint band and a strong band on the same membrane and trust that both are in the linear range without checking multiple exposures — because there's only one "exposure." You scan once, at one intensity setting, and the data is the data.

For quantification, this matters in a concrete way. Suppose your treatment increases target expression 1.5-fold. On a ChemiDoc, if your control band is already at 70% of the detector's saturation ceiling, the treated band gets compressed — your measured fold change might come back as 1.2×. On the Odyssey, both signals sit comfortably in the linear range and you measure the real 1.5×. Gassmann et al. (2009) demonstrated exactly this kind of compression artifact with chemiluminescence and showed that fluorescence-based detection preserved linearity across the same protein range.

Multiplexing and Normalization

This is where the Odyssey has a structural advantage that no amount of clever software can replicate on a chemiluminescent system. With two spectrally separated NIR channels (700 nm and 800 nm), you can probe your target protein and your loading control on the same membrane, in the same scan, without stripping and reprobing. Your target might be in the 800 channel (IRDye 800CW secondary) and vinculin in the 700 channel (IRDye 680RD secondary), and there's essentially zero crosstalk between channels.

On the ChemiDoc, normalization requires either: (a) stripping and reprobing for the loading control — which removes 5–30% of your target signal and introduces variability (Bhatt & bhatt, 2021), or (b) using Bio-Rad's stain-free total protein technology, which is genuinely useful but requires stain-free gels and adds UV-activation steps. Stain-free TPN (total protein normalization) on the ChemiDoc is actually one of its underrated strengths — it sidesteps the well-documented problems with housekeeping proteins changing across conditions (Aldridge et al., 2008) and gives you a lane-level normalization factor before you even do the immunodetection. But it's not multiplexing in the same sense; you still can't image two immunoprobed targets simultaneously.

If you routinely need two-target multiplexing (say, phospho-protein and total protein on the same membrane), the Odyssey is the clear choice. If you're happy with total protein normalization and single-target immunodetection, the ChemiDoc with stain-free gels is a perfectly reasonable workflow.

Sensitivity: ECL Still Has Its Place

One area where chemiluminescence holds an edge — or at least parity — is raw sensitivity for low-abundance targets. Enzymatic amplification (one HRP molecule converts thousands of substrate molecules into photons) means ECL can detect femtogram-level protein quantities. NIR fluorescence has no amplification step; each fluorophore emits a fixed number of photons. For truly scarce targets — think low-copy transcription factors, endogenous signaling proteins in primary cells — ECL on the ChemiDoc can sometimes pull a band out of the noise where the Odyssey shows nothing.

That said, LI-COR has narrowed this gap significantly with brighter dyes (IRDye 800CW is substantially brighter than the old Cy5.5 equivalents) and more sensitive scanner hardware. And there's a subtlety here: a band you can only see with a 10-minute ECL exposure is almost certainly outside the linear range for quantification anyway. If you can't quantify it reliably, detecting it is an academic exercise — useful for a "yes/no, is it there?" experiment, less useful for the densitometry you're probably reading this post for.

Cost, Consumables, and Practical Workflow

Let's talk money and daily annoyance, because these matter.

Upfront cost: A ChemiDoc MP or XRS+ runs roughly $25,000–$45,000 depending on configuration. A LI-COR Odyssey CLx or M is typically $50,000–$80,000. The Odyssey is more expensive, period.

Consumables per blot:

• ChemiDoc: ECL substrate ($0.50–$3 per blot for standard ECL, $5–$15 for femto-grade), standard HRP secondaries ($0.30–$1 per blot). Cheap.
• Odyssey: IRDye-conjugated secondaries are significantly more expensive (~$3–$8 per blot depending on dilution and dual-color usage), but you don't buy substrate. Net per-blot cost is modestly higher.

Hands-on time: ECL requires adding substrate, rushing to the imager, and often capturing multiple exposures to find one in the linear range. If you're sharing the ChemiDoc with six other people and someone's running a time-course, good luck. The Odyssey workflow is: finish your secondary incubation, wash, put the dry or semi-dry membrane on the scanner, press scan. The membrane is stable for days to weeks — you can re-scan next Tuesday if you want. This is genuinely freeing.

Data integrity: Because ECL signal decays, you get one shot (or one series of shots) at your data. If you realize your exposure was wrong, you can re-add substrate, but signal quality degrades with repeated applications. Odyssey membranes are scannable indefinitely. I've re-scanned membranes 6 months later and gotten the same quantification values within 5%.

Software: Image Lab vs Image Studio vs Everything Else

Bio-Rad's Image Lab and LI-COR's Image Studio (now Empiria Studio for newer systems) are both competent densitometry packages. Image Lab has a nice auto-detection feature for lanes and bands, supports stain-free normalization workflows natively, and exports publication-quality images. Image Studio is more spartan but handles dual-channel overlay cleanly and has built-in saturation warnings that flag overexposed pixels — a feature I wish every platform had by default.

Neither software is great at enforcing good quantification practices. They'll happily let you draw ROIs on a saturated band and report a number. The onus is on you to check for saturation, verify your loading control is in range, and confirm linearity. Both export to CSV or clipboard for downstream stats.

Many labs end up in ImageJ/Fiji regardless of what imager they own, either out of habit or because reviewers ask for "ImageJ quantification" specifically (which is not actually a quality standard, but that's another post). The key point: your analysis software matters less than your image quality. A 16-bit TIFF from either system contains the information you need. An 8-bit JPEG screenshot from a phone camera pointed at the ChemiDoc screen does not.

So Which Should You Buy?

If you're doing primarily qualitative westerns (is the protein there or not?), either system works. Buy whichever your department has a service contract for.

If you're doing quantitative westerns — fold-change comparisons across conditions, dose-response curves, anything where you need to put a number on a bar graph and defend it in review — the LI-COR Odyssey is the stronger choice. The wider linear dynamic range, stable signal, and two-color multiplexing solve real problems that make chemiluminescent quantification harder than it needs to be.

If your lab already has a ChemiDoc and budget is tight, you can absolutely do solid quantification. Use signal accumulation mode, always run a loading series to verify linearity for your target, use stain-free total protein normalization instead of GAPDH, and check for saturated pixels before you quantify. These steps close much of the gap.

The worst choice is the one you don't validate. Whichever system you use, run a two-fold dilution series of your lysate, quantify each lane, and plot signal vs. protein loaded. If it's not linear, nothing downstream — no normalization, no statistics, no p-value — will save you.

References

• Aldridge GM, Podrebarac DM, Greenough WT, Bhatt DH. The use of total protein stains as loading controls: an alternative to high-abundance single-protein controls in semi-quantitative immunoblotting. J Neurosci Methods. 2008;172(2):250–254.
• Gassmann M, Grenacher B, Rohde B, Vogel J. Quantifying Western blots: pitfalls of densitometry. Electrophoresis. 2009;30(11):1845–1855.
• Taylor SC, Bhatt DH, Bhatt DM, Posch A. A defined methodology for reliable quantification of Western blot data. Mol Biotechnol. 2014;56(11):966–976.
• Butler TAJ, Paul JW, Chan EC, Smith R, Tolosa JM. Misleading Western blot interpretation: quantitative differences in commonly used housekeeping proteins. BBA Proteins Proteom. 2019;1867(3):308–313.