We got the end of Dr. Davis, and the beginning of Rebuttal by Dr. Brenna.
On the rest of direct testimony, Davis gave us a demonstration of the Optima software on the Isoprime instrument used on the S17 samples, and showed that it was indeed trivial to save the analysis parameters that would document manual corrections. He then demonstrated how manual methods are used, and how seemingly minor changes may result in large reported changes in the delta-delta values.
His cross-examination was unfocused. It started with a fishing expedition for resume errors, a common tactic when there is little else to be found. Young then tried to get him to accept he had ample opportunity to examine data under Botre's supervision at LNDD, and argued back that the amount of data was not realistic. I expect USADA will return to this argument.
When asked if the 5a's weren't still positive, he replied "they aren't even the right peaks, the numbers are meaningless."
Young then tried to trip him up on wording in a declaration regarding reprocessing, and gave up.
On re-direct, Jacobs asked why the reprocessing of the S17 samples on Masslynx was invalid. Davis said that the conversion software did the wrong thing because it was incompletely developed (by him), and inserted data from the Masslynx system. This is part of why the Masslynx data had significant (.5) differences from the Isoprime data. This isn't enough to turn a -6 to a -2, but indicates quantitative problems.
McLaren asked him about "state of the artness" of Masslynx, and seemed satisfied with the answers.
Dr. Brenna was brought back, and appeared to address the peak identification of internal standards by marching through a bunch of sheets showing they all had the same elution time. Comments here wonder if those were fair comparisons, and we'll find out in cross-examination.
He then addressed linearity, and pronounced himself satisfied that peaks in question were in the linear range of the instrument. This dodged some subtle points about low-value shoulder and co-eluting peaks.
He also said that because the instrument correctly resolved the mix-cal's, then it was working right for the samples. Herr Doktor Professor Wolfram Meier-Augenstein had said that the cal mixes did not show the software worked correctly on complicated samples.
There were a number of things he was not invited to testify to, and some careful hedges in places. Expect those to be probed.
We're going to cross Brenna after lunch, then get a brief sur-rebuttal from Davis about linearity, and then be on to closing arguments.
With all the witnesses in, the nice stories are going to come in the closing arguments.