Partial alignment-- first thoughts

Started by infjamc

infjamc Lv 1

First of all, thanks for implementing this tool. However, trying it in real puzzles revealed that using it in practice is more difficult than expected for several reasons:

  1. Right now, there is still no way to align the entire protein in one go by selecting pieces from the different templates.

  2. When doing a partial alignment, the rest of the protein is fixed. This could be problematic when threading across different templates, as their backbone locations could be very different.

A good example of this is #305. For a good overall alignment, the template with a score of 343 under the default arrangement looks pretty good, but the N-terminus needs one of the 243 templates while the C-terminus needs the 250 template. Trying to apply a partial threading to either end, however, will mess up the loops connecting those segments to the rest of the protein. Now, this problem could be solved if there were a way to align the template to part of the protein just like aligning a guide or to move the entire protein in single-piece puzzles as if it were a ligand…

infjamc Lv 1

I've attached two images of what I mentioned in the above post. Also… I have bad news to tell you guys.

…The backbone atom problem is still there!

LennStar Lv 1

I also find this quite annoying.

I would say that a (several) user template in the selection box should available - with drag and drop you could move parts of a template to this artificial one and then make the threading. With this you always know where you are in the threading process from different templates.

Also, if you move the mouse over a part in the template, it should be more visible on the screen where that one is. I think the "note mode" would be good - mouseover number 33 and the note mode is activated for your current Nr. 33.