Drifts in OD to CFU ratios

OD₆₀₀ is often used as a real-time stand-in for plate counts. You prepare different dilutions of the same cultures, measure with OD and CFU pairs, draw a standard curve, and read cell concentration off that curve from then on.

The curve rests on one assumption: that a given OD always means the same number of cells. It does not. The conversion factor drifts as the culture grows. Across one growth curve the OD-to-CFU factor rose 6.6×, from 2.6×10⁸ to 1.7×10⁹ CFU/mL per OD unit.

Read the highlights below. The full story is available in OD calibration factors drift.

The same OD, three different CFU results

An E. coli culture was followed from ~1×10⁶ to 1×10¹⁰ cells/mL, measuring OD₆₀₀ and plate counts side by side. The left chart shows that BactoBox® cells (lavender) is practically superimposed on the CFU data (green). OD axis is adjusted to balance the data points below and over the CFU curve.

Initially, at 2.2 hours, cells are big and effective at scattering light. This leads to relatively few cells per OD; 1 OD₆₀₀ unit corresponds to 2.57×10⁸ CFU/mL. At the end of incubation cells as cells starve, they get smaller and scatter less light. At 11.6 hours 1 OD₆₀₀ unit corresponds to 1.69×10⁹ CFU/mL. That's a factor 6.6× difference.

Same reading, same culture, same instrument. The only difference is when the calibration was made.

Three OD-to-CFU standard curves from one E. coli culture. Each line uses the conversion factor measured at that point in the run. The dots mark the measured OD and plate-count pairs. Reading one OD off all three lines gives answers that span 6.6×.

No single factor fixes it

Averaging does not rescue the method. The geometric mean of the CFU-to-OD ratio across the whole run is 8.59×10⁸ CFU/mL per OD unit. Apply it everywhere and it over-reads by 3.3× at 2.2 hours and under-reads by 2.0× at 11.6 hours. Every constant you could pick lands somewhere in that same trade-off.

Implications

When OD measurements are used to infer bacterial concentration, this assumes a constant calibration factor. It is not. Therefore OD is not a suitable proxy for standardizing actively growing bacterial cultures.

Use BactoBox® instead. BactoBox® measures bacteria directly and does not need calibration factors. BactoBox® demonstrates close to 1:1 agreement with plate counts during the entire growth curve.