1. What were my goals for this week?
- Write a sanity checker for the molecule reader - right atoms/bonds/etc
- Extend Aaron's GNM to elastic and chemical network models
- Look at how to qualitatively analyze motions produced by those models
- Look at how to proceed more specifically on grouping for abstractions
2. What did I accomplish this week?
- Created Sanity (Molecule project) and MotionSeg (MotionSegmenter project)
- Spent about a day learning composition and structure of amino acids
- Got atoms portion of Sanity implemented, bonds portion has correct information but not data structures yet
- Began implementation of MotionSeg using kmeans to cluster motions, plan to use hierarchy to cluster position
- Created beam PDB file to try and get some simple motions working and discovered elNemo hated it
- Got k-means clustering for vector motions working on AKE:
3. Why are numbers 1 and 2 different?
Didn't get bonds/etc done for the sanity checker as the structure for testing these seems unnecessarily complex - need to distill it down to something simpler before I waste CPU cycles on it. Didn't get a chance to look at GNM last week, as Sanity and MotionSeg took up most of my time. Hadn't done any organic chemistry before, so the necessary background to a while to establish. As far as looking to qualitiatively analyze motions and grouping for abstractions, I hope to use the clustering I tried in MotionSeg (see above) and possibly a couple of other different approaches to come up with something that works well for our purposes (might try mean-shift next if I can find a good implementation that works with what I have).
4. What are my goals for next week?
- Try to generate motions on the beam PDB using Aaron's C-alpha program
- Determine if PointList.h is worth using
- Finish Sanity - bonds, etc
- Finish MotionSeg - spatial clustering, different types of motion clustering
5. How does this fit in my bigger picture?
Generate simple motions so we can determine if the clustering algorithm works, which will in turn indicate clusters for abstract motion representation. Sanity makes sure we have what we expect from a PDB and provides a good intro to the chemistry behind this.
6. What did I read this week?
- McMurry, John, Organic Chemistry.