USMA Help

Usher Syndrome Missense Analysis

The Help is presented here as a slide show using s5.
Click here to display a plain page.

Use the arrows of your keyboard to move into the slide show.

Mouse control is also available at the bottom-right corner of the page.

David Baux

CHU-Montpellier

What is USMA?

• Usher Syndrome Missense Analysis
• A software, Website dedicated to the in silico analysis of missense variants,
• related to 9 Usher proteins.
• It uses specific pieces of information,
• alignments, annotations, 3D structures...
• Manually curated.

What is USMA?

USMA is divided into four main parts:

• Otholog alignments
• Domain alignments
• Secondary structure
• 3D structure

Ortholog analysis - Alignment

Conservation of the residue is assessed within a set of ortholog sequences (i.e. the same protein in different species).

Specific features (often unique):

• Window display of the alignement, with a zoom function

Ortholog analysis - AAPI(R)

• AAPI & AAPIR
• Alignment Average Percentage Identity (of the Region)
• Mean conservation of the whole alignement or of a window of 20 residues (Region)
• Available for the position of interest

Ortholog analysis - AAPI(R)

• AAPIRs have also been calculated for the whole alignments
• Window size & positions depends on the size of the protein
• Highlights highly and poorly conserved regions
• Results have been plotted with AAPI and scaled protein pictures (zoom function)

Ortholog analysis - AAPI(R)

Ortholog analysis - Informativity

• P₀ calculated from Greenblatt et al., 2003
• 1-P₀ indicates the probability that a invariant position is actually functionally constrained

Ortholog analysis - Phylogeny

• A phylogenetic tree has been built for each alignment

Ortholog/Domain analysis - Venn diagram

• Alternative representation of the alignment at a given position,
• taking physico-chemical properties into account (SVG format)
• Color code

Ortholog/Domain analysis - Venn diagram

• green: wild-type residue
• red: mutant residue not found in the alignment (at the given position)
• purple: mutant residue found in the alignment
• orange: other residues found in the alignment
• grey: other residues not found in the alignment

Ortholog/Domain analysis - Venn diagram

• The size of each letter depends on the representativity of the residue at the position

Ortholog/Domain analysis - WebLogos

• WebLogos are also available for the region (21 residues)

Domain analysis - Alignment

• Alignments all come from pfam or prosite
• Include up to 2000 sequences (usherin fnIII)
• Useful to identify structurally crucial residues
• Be careful with less conserved residues, they can have a specific role in your protein!!

Domain analysis - Alignment

• Localization of the domain and summary table of the analysis

Domain analysis - Alignment

• As for orthologs, real-time svg Venn Diagram and WebLogo
• Possibility to display (clustal format) the complete alignment...
• And therefore to download it for further analysis

Secondary structure - Psipred

• Psipred predictions
• Based on ortholog alignments (better prediction than with a single sequence)
• 3 state prediction: Helix, Strand, Loop

Secondary structure - Frequencies

• Frequencies of observation of each 20 residues in helices and strands...
• Have been estimated based on 8,365 3D structures, representing 1,598,587 residues
• Inclusion:
  • non-redundant structures (14,550 structures at first), extracted from the PDB
  • Annotated representative structures presenting a resolution < 2.5 Å

Secondary structure - Frequencies

• Exclusion:
  • Helices of less or equal to 3 residues
  • Strands of less or equal to 2 residues
• Graphs of these calculations available, mapped with:
  • Chou-Fasmann, 1978
  • Creighton, 1983
  • Costantini, 2006
• Frequencies used to determine wether WT and M residues are favorable or not to their predicted stII

Secondary structure - Frequencies

3D analysis - WT structures

USMA is able to analyse 3D structures

• WT structures are "manually" generated, using Domain fishing or @TOME2 for template recognition
• Modeller for model construction
• SCWRL4 for side-chain refinement
• STRIDE for secondary structure assignment
• Molprobity for assessment
• We try to select templates with at least 30% identity (sometimes a little lower)

3D analysis - WT structures

• 3D coverage graphs of each protein are available

3D analysis - Mutant structures

• USMA generates real-time mutant structures using Modeller
• USMA's "3D engine" computes:
  • involvement of wild-type residue
  • involvement of mutant residue
  • differences

3D analysis

• Involvement means predicted...
  • SAS (Solvent Accessibility Surface), computed with msms
  • Hydrogen bonds (calulation including distances and bond angles)
  • salt bridges network
  • disulfide bridges
  • steric clashes

3D analysis

USMA also interprets secondary structures features such as amino acids propensities for specific N-Cap, N1-3, interior, C3-C1 and C-Cap positions of α helices and i,i+3 and i,i+4 possible interactions

3D analysis

• Results are rendered as a summary table...

3D analysis

• And real-time static pictures made with Pymol...

3D analysis

• But also 2 fully functional Jmol Applets

Then the user can perform its own analysis of the models

Additional ressources

• In addition, USMA provides several information:
  • UNIPROT annotations concerning the variant if available
  • Direct link to LOVD-USHbases
  • link to NCBI's graphical view of the region

Last slide

• USMA's just in silico!
• Do not forget other sources
• DNA controls, familial segregation, splicing, and, of course,
• LOVD-USHbases
• Thank you!!