3 ms·

Not properly blanking your UV/Vis measurement is an amateur mistake, that really should not happen in this kind of setting. UV/Vis spectroscopy is a rather unsp

Not properly blanking your UV/Vis measurement is an amateur mistake, that really should not happen in this kind of setting. UV/Vis spectroscopy is a rather unspecific method, many things can absorb the same wavelengths. You need to have a very good idea of what is in your sample to be sure that you're measuring the right thing. In this case there are two things that absorb at the same wavelength, the vaccine and the polysorbate. The proper reference measurement would be all components except the vaccine, which you then compare to the full vaccine. It reads a bit like the used water or another buffer as blank here, which is really inexcusable.

_Wintermute 6y ago
What amazes me more is when they got this huge discrepancy they didn't confirm it with qPCR, or contact the supplier who would likely point out the polysorbate. You must really feel for AZ at this point, and wonder how much they want to partner with academic scientists at such a critical stage.
- Closi 6y ago
  > You must really feel for AZ at this point, and wonder how much they want to partner with academic scientists at such a critical stage. Considering this looks like the most promising vaccine candidate for large-scale global rollout and has by far the most orders, I'm sure they are still happy they are working with the world-class Oxford vaccine team.
afarrell 6y ago
> amateur mistake You wouldn't believe how many software engineers with years and years of experience make amateur mistakes during interviews. Pressure does weird things to people's performance.
- fabian2k 6y ago
  That kind of decision must have involved a whole bunch of people. And I really can’t understand why you would trust the UV/Vis data over qPCR in this case. An additional component absorbing at the same wavelength the first thing you would suspect if you see a difference in this direction.
hedora 6y ago
This is a classic communications breakdown. One group mixed the vaccine, and the other tested. The first group added a filler that throws off the UV measurement. They didn’t tell the second group about the additive because they didn’t realize the second group would use UV to measure the concentration. Neither group could have avoided the problem without access to information they did not have. Both groups under-communicated. This sort of screwup is by far the leading cause of software bugs (most bugs occur at the interfaces between teams).

_Wintermute 6y ago

What amazes me more is when they got this huge discrepancy they didn't confirm it with qPCR, or contact the supplier who would likely point out the polysorbate. You must really feel for AZ at this point, and wonder how much they want to partner with academic scientists at such a critical stage.

Closi 6y ago
> You must really feel for AZ at this point, and wonder how much they want to partner with academic scientists at such a critical stage. Considering this looks like the most promising vaccine candidate for large-scale global rollout and has by far the most orders, I'm sure they are still happy they are working with the world-class Oxford vaccine team.

afarrell 6y ago

> amateur mistake You wouldn't believe how many software engineers with years and years of experience make amateur mistakes during interviews. Pressure does weird things to people's performance.

fabian2k 6y ago
That kind of decision must have involved a whole bunch of people. And I really can’t understand why you would trust the UV/Vis data over qPCR in this case. An additional component absorbing at the same wavelength the first thing you would suspect if you see a difference in this direction.

hedora 6y ago

This is a classic communications breakdown. One group mixed the vaccine, and the other tested. The first group added a filler that throws off the UV measurement. They didn’t tell the second group about the additive because they didn’t realize the second group would use UV to measure the concentration. Neither group could have avoided the problem without access to information they did not have. Both groups under-communicated. This sort of screwup is by far the leading cause of software bugs (most bugs occur at the interfaces between teams).