Michael Schneider (Talk | contribs)
(New page: == Thursday 03/13/08 == Not so much done today ... Finally I moved to my new desk ^^. Ive collected the data of the last 5 structure pairs, so far no obvious errors so I can start the...)
Next diff →
Not so much done today ...
Finally I moved to my new desk ^^.
Ive collected the data of the last 5 structure pairs, so far no obvious errors so I can start the global analysis. I guess a good plan is just to summarize the data conclusion in the easiest possible manner e.g. just say how many structure pairs where there, which had better energies for at least 1 NMR structure in repack/design.
I also made some scatter plots which seem to be a good way to easy show the general trends of the entire testset. For the energies, I still have to normalize the data and exclude outlier. Maybe I should just trash a certain upper quantile or so.
Also, the investigators of PSVS already investigated, that minimization of NMR structures can lead to a significant increase of their quality measures. It would be interesting to check, whether these measures correlate with the design performance. Computing the z-scores for my structures after minimization could indeed be useful and interesting!!! But first, I should update my data "book" just to have the collection of data present\
I decided to give RefWorks a try and finally start writing the paper ... At least Methods section were I can test the usage of RefWorks to decide whether it works for me.