Secondary Structure Pattern Selection

[trudack]

The analysis tool for the abinitio and insertion parts should also provide the possibility to select individual secondary structure patterns. Right now it just gives the average pattern.

For the structures in this example it would be useful to select a structure from the predicted structures that has a helix from resid 292 to 305 or the structures that are closes to it.

Probably an input list {"ss element, resid X to Y" "ss element, resid A to B" } would be reasonable.

Resid 292: H: 100.00 T: 0.00 C: 0.00 G: 0.00 E: 0.00 B: 0.00 I: 0.00
Resid 293: H: 100.00 T: 0.00 C: 0.00 G: 0.00 E: 0.00 B: 0.00 I: 0.00
Resid 294: H: 100.00 T: 0.00 C: 0.00 G: 0.00 E: 0.00 B: 0.00 I: 0.00
Resid 295: H: 100.00 T: 0.00 C: 0.00 G: 0.00 E: 0.00 B: 0.00 I: 0.00
Resid 296: H: 100.00 T: 0.00 C: 0.00 G: 0.00 E: 0.00 B: 0.00 I: 0.00
Resid 297: H: 95.00 T: 5.00 C: 0.00 G: 0.00 E: 0.00 B: 0.00 I: 0.00
Resid 298: H: 80.00 T: 10.00 C: 10.00 G: 0.00 E: 0.00 B: 0.00 I: 0.00
Resid 299: H: 60.00 T: 10.00 C: 30.00 G: 0.00 E: 0.00 B: 0.00 I: 0.00
Resid 300: H: 30.00 T: 30.00 C: 40.00 G: 0.00 E: 0.00 B: 0.00 I: 0.00
Resid 301: H: 35.00 T: 45.00 C: 20.00 G: 0.00 E: 0.00 B: 0.00 I: 0.00
Resid 302: H: 35.00 T: 45.00 C: 20.00 G: 0.00 E: 0.00 B: 0.00 I: 0.00
Resid 303: H: 35.00 T: 40.00 C: 25.00 G: 0.00 E: 0.00 B: 0.00 I: 0.00
Resid 304: H: 35.00 T: 20.00 C: 45.00 G: 0.00 E: 0.00 B: 0.00 I: 0.00
Resid 305: H: 20.00 T: 15.00 C: 65.00 G: 0.00 E: 0.00 B: 0.00 I: 0.00
Resid 306: H: 0.00 T: 0.00 C: 100.00 G: 0.00 E: 0.00 B: 0.00 I: 0.00
Average sequence : H H H H H H H H H H H H H H H H C T T T C C C
final results (frames with highest scores - 89) : 9
number of hits: 1

I will sent some files to test to @ryanmcgreevy by mail

[trudack]

The secondary structure tool needs to be optimized as a whole

• solve issue with empty summary text file for some structures
• individual search pattern for all secondary structure element types with resid selection commands (as described above)
• adapt ss and clustering model maker selection to VMD atom selections
• clean up the scripts, there are a lot of non used options or out commented lines in the code for ss analysis
• internal VMD plot solution without gnuplot
• double check that the scoring to find the average structure is correct
• as soon as the Modeller part works we can use the Modeller secondary structure restraints to generate exactly the average structure

[ryanmcgreevy]

Just updating this issue with the contents of a recent email from @trudack on the subject:

I encountered one key issue, that the secondary structure analysis does not work for some files. The summary text file has no percentage analysis so this part seems to fail. I do not see the difference why sometimes it fails and sometimes not.

I have put one working example (rpn11) and one not working example (helix1_2) with all scripts and files here:

https://www.dropbox.com/sh/ij53ohseisn7ex9/AAAWp_ypVbmsNZAi9JtDH7aNa?dl=0

I have more not working files if you need more examples but right now I have no clue why it is not working and we really should fix this.

[trudack]

I was stupid, and Joao is right it is just the wrong chain ID in the script. After fixing that full length model keeps the chain IDs I forgot to adapt it in the analysis script. Sorry. However, this brings up the point again that there should be an obvious warning whenever an atom selection is wrong or empty. This would make it much easier to find identify errors for newbies to VMD.