What you need to know about OD600

on 8th of December, 2008 in Cloning & Expression
About the Author:
Nick Oswald started Bitesize Bio on a Macbook on his kitchen table in 2007 while in his 7th year of working as a molecular biologist in biotech. He made it his day job in 2010 and has been loving it ever since.

If you use a spec to measure cell density, you may making a very common mistake and taking inaccurate measurements as a result.

Specs are often used for measuring the density of suspension cultures, but the mistake that many people make is to record the OD given by the spec as an absolute value.

The OD value represents the amount of light that is absorbed by your sample. But that value is affected by the intensity of the light beam in the spec, and the spec design. This means that similar samples will give completely different OD values in different specs due to the specs having different bulbs, or even in the same spec over time, as the beam intensity reduces with the age of the bulb.

So recording an OD value in your lab book does not really mean anything as this number is as much dependent on your spec as the density of your culture.

What you really want to know

What you really want to know from an OD600 reading is the density of the cells e.g. in cells/mL. And to get this you need a standard curve. In other words, like any other spec-based experiment you will ever perform,  you need to calibrate the absorbance value against the number you actually want to know.

For some reason, people don't seem to remember this for OD600. I can't think of any other experiment where people record the absorbance value as an absolute number like they do for OD600, but maybe you can think of some – tell me in the comments :).

Constructing a standard curve for OD600 is a bit tedious, but it is a good exercise to go through every 6 months or so (and for each spec you use).

Calibrating your OD600 measurements

Start by making a suspension culture of the cells you are interested in, then diluting to obtain a series of samples with ODs of =2, 1, 0.8, 0.6, 0.4, 0.2 and 0.1 on your spec. (NB: don't make serial dilutions as these are very inaccurate).

For each OD you then want to know the cell density in cells/mL. So for each OD, make dilutions of 1 in 1×10^7, 1×10^6 and 1×10^5, then plate 1mL of each onto suitable plates and grow them up. Then count the number of colonies formed on the dilution that gives the most appropriate number and multiply up by the dilution factor to obtain the number of cells/mL in the original sample. These values can then be used to construct a calibration curve of OD vs cells/ML.

The conversion factor for your spec (the number of cells/mL represented by 1 OD unit) will be equal to the gradient of the linear portion of the curve (normally up to about OD=1).You can now use this factor to convert your OD reading to cells/mL and as long as you calibrate whatever spec you are using in the future, this number will be absolute and comparable between experiments, specs and years.

Just when you thought it all made sense…

A word of warning though, since the OD of a sample is dependent on the size and shape of the particles in it, different cell lines can have completely different relationships between OD and cells/mL. This means that a separate calibration will be needed for each cell type you use, which is tedious, but better than recording meaningless and arbitrary numbers in your lab book.

15 thoughts on “What you need to know about OD600”

  1. wendy says:

    You are absolutely correct when you say that this is one of the simplest, yet most misused techniques. The number one reason that plasmid isolation kits do not seem to work is due to the amount and quality of the starting culture. Long ago, kill curves were an integral part of undergrad teaching labs and part of all good work with bacteria. Doing it right could take several days, however, so this may be why it fell out of favour. The important thing that one sees when doing a growth curve over time is the plateau. Dead cells have an OD (but won't form a colony); It is important that cells be harvested in Log Phase to avoid the dead cells. Furthermore, do not dilute an overgrown culture to the right OD and then spin down the cells (many may be dead). I have had to explain this to (experienced?) techs and post-docs as well as new grad students.

  2. Nick says:

    Thanks for that Wendy – some good points there.

  3. Aaron Stephan says:

    "I can't think of any other experiment where people record the absorbance value as an absolute number like they do for OD600, but maybe you can think of some – tell me in the comments :)."

    Thanks for the article. Very interesting. Maybe this demonstrates my ignorance, but we use the absolute value of A260 all the time to measure DNA concentrations without first establishing a standard curve. I understand that the cuvette pathlength and extinction coefficient are both required, but in a sense, how is that different from the OD600, where the cuvette is still generally 1cm. Yes, I understand the absorbance can very between cell lines, but I don't understand why it can vary between spectrophotometers.

  4. Nick says:

    Hi Aaron,

    To measure a DNA concentration using A260 you need to determine the relationship between the absorbance and the DNA concentration… which can only be done with a calibration curve.

    I would be surprised if you normally record the A260 value in your lab book as the "concentration" of a DNA sample (e.g. the conc of a plasmid prep as 0.5 absorbance units). More likely you record it as ug/ml or something similar – so there has to be a calibration curve in there somewhere.

    If you use a spec that is made specifically for measuring DNA concs, such as a biophotometer, the standard curve is normally pre-programmed and the conversion from AU to conc is done automatically by the machine so maybe that is where the confusion arises.

    OD600 varies between machines because, unlike A260 it does not measure molecular absorption, but light scatter. The amount of scatter is profoundly affected by various aspects of the spec, including the distance between the cell holder and instrument exit slit, geometry of the slit, the monochromator optics and bulb intensity. This is why different specs give different OD600 measurements.

  5. Aaron Stephan says:

    I usually write my A260 down and then calculate the [DNA] myself using Beer's law. We have two spectrophotometers (a cuvette-based and a nanodrop), both of which just caculate the [DNA] the same way I do: c=A/(EL), where E is the extinction coefficient, and L is the pathlength.

    I do understand that at some point, the extinction coefficient for DNA was determined empirically by making a standard curve. But I do not / can not believe that the A is a completely arbitrary value that can be completely different between different spectrophotometers. A is a definable term based on the percent of light that is emitted and then detected. Since you blank your sample before you start, you define this as 100% transmittance, which is an A of 0. I think no matter how bright the light, the substance that you are measuring essentially acts as a filter (or in the case of cells, scatter). Therefore, if the substance filters/scatters 50% of the light, then it will filter 50% of the light regardless of the intensity of the light (within certain limits of course). If this is not true, then Beer's law is not true. It would also have to take into account the intensity of the light source.

  6. Aaron Stephan says:

    AU is arbitrary units, which is generally used for Fluorescence. The fluorescence is arbitrary because it does depend on the intensity of the activating light.

    A is absorbance, which is a defined, repeatable unit

  7. Aaron Stephan says:

    Sorry, my thumb hit submit before I was finished.

    AU is arbitrary units, which is generally used for Fluorescence. The fluorescence is arbitrary because it does depend on the intensity of the activating light.

    A is absorbance, which is a defined, repeatable unit of measurement, which is defined by the amount of light that is transmitted:

    A = log10 100 / %T

    I do understand that for the most accurate measurements, you would want to do a standard curve because each measurement tool has its systematic error. However, to suggest that A is arbitrary is like defining meters or degrees celcius as arbitrary, too.

  8. Nick says:

    Hi Aaron,

    Hmmm… maybe I should have refreshed my mind by a text book before writing this one! You are right about the absorbance being absolute for molecules such as DNA, so I should not have extrapolated the problems with OD600 to DNA and other spec-based assays.

    However, because light is scattered by cells in suspension, the spectroscopic measurement of cell density is a different situation.

    Unlike molar absorption, the scattering of light is a more complex optical measurement, which is affected by the setup of the individual spec. More info on this is available here: http://www.microbiol.org/white.papers/WP.OD.htm

  9. Aaron Stephan says:

    Sorry for leading your point off track. I looked at the website about the complexity of light scattering, and I can understand how the light scatter could be affected by the intensity of the light beam, the shape of the slit, and other factors specific to each spectrophotometer.

    I vaguely remember a former professor telling me that because you are measuring different things (absorbance for nucleic acids and proteins and scatter for cell turbidity), that's why we use different prefixes (A260 vs OD600).

  10. Nick says:

    Thanks Aaron, your points are a great addition to this article!

  11. DK says:

    This means that similar samples will give completely different OD values in different specs due to the specs having different bulbs, or even in the same spec over time, as the beam intensity reduces with the age of the bulb.

    OD600 is certainly a crap value that shouldn't be used/published the way it is but your explanation as to why it is so is completely incorrect. It has nothing to do with the bulb/monochromator/etc – for that, there is blank value that should be taken and an internal calibration that ensures linear response with respect to the incoming light. Once these two are done, all spectrophotometers will give the same reading for *absorbing* samples. But what we measure with OD600 is not absorbance, it is scattering!

    So here is a correct explanation: OD600 as a measure of cell concentration relies on scattering of the light – some light is "lost" because scattered light now travels in all directions. So what gets to the photomultiplier has two components: A1+A2 where A1 is the amount of light that was not scattered (function of cell concentration) and A2 is some proportion of scattered light that still made it in (scattered light coming from a slit in the cuvette holder is roughly a cone of light). It is A2 that is variable from spec to spec since it depends on geometry of the spec: sizes of the slit in the cuvette holder and photomultiplier, and also distance from the cuvette to the photomultiplier (the closer it is, the more scattered light will be mistakenly recorded as non-scattered thus lowering OD600 value).

    But yes, of course, when it matters, one should take calibration curve – not only OD600 is not a linear function of cell concentration but different cells scatter differently. (For example, cells with inclusion bodies seem to scatter considerably more; different strains/species may scatter differently, etc).

  12. Pingback: Homepage
  13. Avatar of wensan0000 wensan0000 says:

    Very impressive article. I am new hand in microbiology and would like to ask the calibration for OD600 measurement. Will the results of calibration vary between different stages of growth? ( eg. Will ythe value of CFU/ml in lag or log phase be different when OD600 value is adjust to the same?) Thank you.

    1. Avatar of JackBean JackBean says:

      yes, it will be different, because in the stationary phase some of the cells will be dead already, as is written in the article. Thus the OD600 will be higher, but CFU will be lower.

  14. Pingback: URL

Leave a Reply

Poll!

What sector do you work in?