search

moltimate / moltimate-frontend

Moltimate's frontend source code.
GNU General Public License v2.0
8 stars 2 forks source link

Motif Maker with three new motifs: 1e3v, 5icd, 1ds2 #102

Open paulcraig opened 3 years ago

paulcraig commented 3 years ago

I am testing the motif maker in http://moltimate.us-east-1.elasticbeanstalk.com/. I selected three new proteins with motifs in the M-CSA that may be not be in Moltimate already. I tested them against self, homologs and random.

  1. 1ds2, streptogrisin B, EC 3.4.21.81, 6 residues in the active site
  2. 5icd, isocitrate dehydrogenase, EC 1.1.1.42, 4 residues in the active site
  3. 1e3v, delta 5-3-ketosteroid isomerase, EC 5.3.3.1, 4 residues in the active site

Expected Behavior

  1. In self test, the created motif should find all residues as expected. Clicking on that alignment should bring up the residues and a report in the box in the upper right hand corner.
  2. In homolog test, the created motif should yield an alignment report with every PDB entry with the same EC class. The alignments should report either "n/a" or an rmsd value in the left hand side. Clicking on the results with an rmsd value should display the alignment and the aligned residues. Clicking on a non-aligned structure (n/a) should result in a message that no alignment was found. Failure to complete the PDB id and EC Class fields should give an error message "You must enter the PDB id and EC class to create the motif."
  3. In random test, the user selects the number of random structures to test (default is 1). It should result in an "n/a" value for the alignment.

Current Behavior

  1. Self tests for all three were successful and quite quick (< 30 s). When I clicked on the listing to see the alignment, the view was as expected. SUCCESSFUL
  2. Homolog tests for all three were quicker than expected (<60 s) and provided the expected alignment report. Clicking on the results with an rmsd value displayed the alignment and the aligned residues. Clicking on a non-aligned structure (n/a) resulted in a blank white page. Moltimate had to be reloaded. Failure to complete the PDB id and EC Class fields gave error message "! [object Object]" in orange.
  3. Random test looked at one alignment and reported n/a. Clicking on that alignment crashed Moltimate again. Also, when I tried to change the number of random structures to test, Moltimate crashed.

Possible Solution

  1. Whenever there is an alignment of n/a, disable the alignment listing and make it non-clickable

Steps to Reproduce

  1. enter the PDB id, ec class and residues in motif maker (see the list provided below)
  2. test for self, but omit the PDB id and EC class (to get the object Object error)
  3. test for homologs and click on one of the n/a alignments
  4. test for random and try to change the number of random structures to align.

Context (Environment)

Motifs were created and tested, but crashed the system quickly.

The use of motif maker is to create new motifs that you wish to use to query structures against. The current design of Moltimate does not support this yet and it is beyond the scope of the current project to make this change, so I have added it to the Future project ideas list below. It would still be very helpful to remove the errors and crashing behavior, however.

Moltimate crashes during alignments when a new motif is created with Motif Maker. This happens when you test homologs or random structures and click on an alignment with an rmsd value of "n/a". It also crashes when you try to change the number or random structures to be tested.

Detailed Description

Here are the details for each of the motifs I created.

PDB id: 1e3v. EC class: 5.3.3.1 Asp A 40 Tyr A 16 Asp A 100 Asp A 103

PDB id: 1ds2. EC class: 3.4.21.81 Ser A 214 Ser A 195 Gly A 193 Asp A 102 His A 57

PDB id: 5icd. EC class 1.1.1.42 Tyr A 160 Asp A 307 Asp A 283 Lys A 230

Future project ideas

  1. The results of the homolog test for created motifs should be sortable by rmsd.
  2. User-created motifs should be stored in a separate library that the user can choose to (a) add to the regular search or (b) use for a limited search.
  3. When homolog testing is done, a report should be available on the number of structures tested, the results for each individual structure and a summary of results (number that did not align, number that did align and the range of rmsd values).
  4. When random structure testing is done with motif maker, the default should be 10 structures, with a similar report as for homolog testing.
  5. It occurs to me that we can't use these new motifs to screen structures of unknown function. Item 2 from this list would enable us to do this.