ldomic / lintools Goto Github PK

The domains at the moment can only be specified as range of residues.

• Allow to select by groups of residues e.g. [1,3-6,20]
• Option for user selected colorscheme
• More ways of showing things - e.g. dashed circle for something like binding site

In occurrence analysis clock diagrams the total length of interactions is shown, however, nothing is expressed about the character of the interaction - i.e. on/off rate

• decide on the way the rate is going to be displayed

The ligand selection in the start allows to choose a ligand residue only by resname, which can cause issues if there are several ligands with the same name. Therefore, an additional factors of segid and residue number should be included

I noticed that the configuration file does not get read properly, since I have changed many options. It is time to debug and update the configuration file by adding new features and removing the old ones.

Check user input before asking diagram type, and then offer the disgrams available according to the input

If the figure contains only a few residues or they are all on one side of the ligand, the function for measuring the size of the svg figure fails.

Should the area molecule is drawn in depend on the size of the molecule?

• Check the values RMSF produces
• change the colorbar so that red shows most movement
• allow to run occurrence analysis of multiple trajectories at the same time as RMSF (average RMSF over the multiple trajectories)
• output the numerical values of RMSF (either in config or provide a flag that will print them)

Legends for:

• Amino acid types
• Domains
• Hydrogen bonds, salt bridges
• RMSF results
• Trajectory identities
• Additionally, need to fix hydrogen bond representation in domain graphs

• look for bridging waters
• give ions, lipids and sugars representation in the diagram and analysis

At the moment a mol2 file of the ligand is required. This requirement could be removed by including a software like OpenBabel in the Dockerfile and making a ligand mol2 file as part of the LINTools run.

In case openbabel misbehaves

Need to find a way to make a nicer 2D rendering of the molecule using SMILES instead of mol2 and not loosing track of the atom identities.

Enhance residue placement using more than one atom to determine place on the hull, instead using the 3 closest ones or 2 closest neighbours to the closest ligand atom.
The distance from the hull can be chosen by either sum of all distances or the single closest distance..

Another idea - introduce weighting parameter in this function

For some reason, when LINTools outputs the drawn clock analysis, it output's it to a location 2 directories up from the directory from which it was launched.

Have trajectories with different run lengths to plot appropriately

At the moment spelling the diagram type wrong causes problems.

• Make diagram type selectable
• Mix & match diagrams (Clock+amino, domains+amino)

Sometimes MDAnalysis gives a bug when counting hydrogen bounds by type. I plan to circumvent this problem by writing my own function to deal with this.