Going for the longest post ever prize here......
RTMcFadden said:
For simplicities sake, if you want to continue this discussion, let’s agree on some terminology. Let’s equate signal strength with absorbance, and refer to “spiked” and “unspiked” as reference or control samples (positive and negative). Also, instead of the term ”tester,” let’s call the people who collect the samples “collectors” and the people to run the analytical tests “analysts.”
+1....and can we add "athlete sample" for a sample that is being tested for rEPO, on the basis of the criteria established by analyzing the control and reference samples... this is where the wires are crossed this time...(last time it was because I misunderstood 'adjusting the sample' to mean adjusting the original (e.g.120ml) sample, when you actually meant adjusting the sample of retentate).
So, explaining slowly to avoid further mix-ups, and hopefully facilitate short answers, here goes:
RTMcFadden said:
I don’t know exactly how they obtain their positive and negative reference samples (external standards).
I wasn't asking about how the reference standards are obtained (the Vrijman report indicates there were several methods, one of which involves control and reference samples.) I was asking about what happens to athletes' samples in routine doping analysis. The links explaining the standard process of analyzing athletes samples indicate this is done
without adding any rEPO analyte directly to that athletes sample. Instead, analysis steps 1 - 4 are performed on the sample and the results compared with the results of control and reference samples. Is this correct?
Clearly, adding different amounts of analyte to a sample and performing the sample analysis after each addition would allow for a more accurate measure of absolute concentrations.....but the risk of stuff-ups is serious. I just want to confirm that you don't think this type of process is involved in analyzing the athletes' samples, because from one of your previous posts I
wondered if you did....and I would hate for us to be talking at cross purposes
RTMcFadden said:
I would think it's either because it would wash out the detector or because the band would be so big as to run (bleed) into other bands.
OK, we are both talking about some kind of sample adjustment to improve visualization. So that answers the central question about whether, in your view, rEPO analyte formed any part of that adjustment.
RTMcFadden said:
Here’s what I have a problem with. In the discussion regarding the probability of correctly selecting LA’s samples, it was noted by somebody that the testing took place over 20 days.....it’s entirely possible that one person could have performed all of this testing....
Collection took place over 20 days, i.e. 20 stages of the TDF. I have never seen any information about how many days the actual analysis took. Information about how many people were involved in analyzing the samples, and how much uninterrupted lab time they had to mess around with anything might be interesting. As a way of investigating how likely it is the samples were deliberately spiked, that information would be more useful than exploring the technical possibilities. As a way of learning more about the testing process, pondering the technical issues is more useful............
Of course, any discussion about malicious spiking is in the context that it's unlikely someone, with unknown motive, managed to break the study blind. So there's no need for the audience to get quite so excited!
RTMcFadden said:
What bothers me is when the results are 100%, was there no nEPO, was it below the detection limit, what’s going on. This is why I’m more interested in the concentration.
I would certainly buy the human error explanation for that single result; and agree that info on the concentration of nEPO would help clarify the matter. I don't buy the idea of six incidents of human error all happening randomly to one athlete - but then who would?
RTMcFadden said:
Comparison? ...The data indicates that the lower LoQ is 125 UI/L and that the upper LoQ is less than 1470 UI/L but more than 906 UI/L for rEPO.
Not sure which of my comments you are responding to here, but that paragraph did confirm what I was thinking about the IEF process, so I appreciate you taking the time. I really think those data above are tEPO not rEPO. That's how Ashenden describes them in the text associated with the table that I quoted previously.
RTMcFadden said:
Some information regarding the relationship between nEPO and rEPO can be found .....Ashenden indicated that a paradigm exists “which holds that endogenous erythropoietin production is suppressed when the red cell mass has been increased beyond the homeostatic set point.” .....So, the addition of rEPO would raise Hb concentration, thereby decreasing nEPO concentration, which makes physiological sense.
That makes sense to me too. It's one of the reasons why I think the 'positive test' criteria include the ratio of rEPO to nEPO....even though rEPO is expressed as a fraction of tEPO. From Ashenden's data and some stuff I read about altitude effects, I suspect there is quite high natural variability in both rEPO and nEPO. Such variability would make these data difficult to interpret with sufficient certainty to impose sanctions, and the rEPO:tEPO criterion would be more useful. I was actually not asking about fluctuations in rEPO and nEPO within the body.
What I want to know is whether the different forms of EPO are affected differently by any of steps 1 - 4 of the analysis. My very obscure signal processing point is relevant to the EPO analysis procedure, if and only if the answer to that question is NO.
Here's the signal processing point in gory detail:
Lets consider a sample containing differing amounts of A, B and C...and express it as SAMPLE = xA + yB +zC + solution. Then we filter and blot the sample and are left with whatever doesn't get lost in processing...lets call that FINAL = f1xA + f2yB + f3zC (where f1, f2 and f3 are the fraction of A, B and C which survive steps 1 - 4.)
If A, B and C are quite similar, then it seems likely that f1 = f2 = f3 (approximately).
If one can accurately measure the ratios, then one can accurately state A, as a percentage of SAMPLE, without knowing anything about the absolute value of SAMPLE......so
if the answer to the above question is NO, it might be possible to know 'the %' more accurately than the absolute concentrations, which are limited by the error in f123 (+/- 6%?).
So that would be another good reasons to include the ratio of rEPO to nEPO in the 'positive test' criteria.......and it would make it pretty tricky to calculate the absolute quantity of rEPO required to maliciously spike a sample.