185
edits
Changes
Jump to navigation
Jump to search
mLine 103:
Line 103: + + + + +
no edit summary
::::::::“Not-C” does not support “the paper says “Not-C”” and the circularity is broken.--[[User:Toffeeman|Toffeeman]] 09:07, 1 August 2008 (EDT)
::::::::“Not-C” does not support “the paper says “Not-C”” and the circularity is broken.--[[User:Toffeeman|Toffeeman]] 09:07, 1 August 2008 (EDT)
:::::::::In the third experiment, additional clones were used. They write: "Clones were chosen to maximize the diversity of colony morphologies from each sample, in case only certain types could produce Cit+ mutants." This is different from the selection of clones in the first and second experiments. Also in contrast to the second experiment which grew cells in DM500 media, they used DM1000 and DM2000 media to grow the cells in the third replay. Cells were plated at higher densities in the third replay experiment.
:::::::::As the authors comment, when they designed the third replay conditions, they assumed that their changes would not cause significant differences. Still, there clearly were differences in the results and there were some differences in the conditions. Note that if, as Andy initially suggested, the initial cultures were contaminated with Cit+ cells, one might expect that contamination to scale as well between the second and third experiments.
:::::::::In any case, the authors clearly noted the anomaly, admitted that they're not sure why it happened and cited the statistics from the third replay experiment in their paper (The results still trended with the previous experiments). It's up for readers to decide whether their conclusions are reasonable, based on the data the authors have presented. Most likely, future experiments, particularly with whole genome sequencing of the isolated clones, will confirm or deny the paper's conclusions. That's pretty much how science works at the cutting edge.--[[User:Argon|Argon]] 12:14, 1 August 2008 (EDT)
:3)The claim that the use of Monte Carlo resampling test was an error is unsupported by argument or counterclaim. ''Why'' is the use of that test an error? Problematically, if the test itself is invalid, why are you willing to accept the p-value for the third replay? Furthermore, the Monte Carlo test was not the test used to combine the three experiments to arrive at the final p-value of p<0.0001 - that was a Z-transformation method. Perhaps your real objection lies with this technique rather than the Monte Carlo test.
:3)The claim that the use of Monte Carlo resampling test was an error is unsupported by argument or counterclaim. ''Why'' is the use of that test an error? Problematically, if the test itself is invalid, why are you willing to accept the p-value for the third replay? Furthermore, the Monte Carlo test was not the test used to combine the three experiments to arrive at the final p-value of p<0.0001 - that was a Z-transformation method. Perhaps your real objection lies with this technique rather than the Monte Carlo test.