Explore sequence, structure,
and function relationships

How to use MODexplorer

PDF version of this help PDF icon

Overview

MODexplorer is an integrative web server for:

It takes as input:

In the output it enables to:

Submitting the job

MODexplorer accepts as input either the structure or the sequence of a protein.
You can select the input type by using the 'Choose input type' field.

protein_input

Protein structure input mode

A protein structure can be submitted to MODexplorer in two ways:

In both cases you are optionally asked to specify the chain.
pdb chain submission field
Notice that if you leave chain id field empty, MODexplorer will consider the first protein chain found in the PDB. However if you specify a chain that is not present in the PDB, this will be considered an error.

Protein sequence input mode

The sequence can be pasted or uploaded. Both FASTA and raw sequence formats are accepted.

Example of inputs:

HHblits and HHSearch parameters customization

MODexplorer allows customizing several paramaters of the embedded HHblits and HHSearch programs.

For a detailed explanation of each HHblits parameter please refer to:

As for the HHSearch ones:

Multiple sequence alignment (MSA) of the query family

You can optionally submit your own MSA of the query family in FASTA, CLUSTAL, PHYLIP, or MSF format. Click on "Advanced options - OPTIONAL". The following form should appear: User MSA submission field After you upload a file, two required additional fields will appear, TARGET sequence name in MSA and Alignment format: Required fields for MSA submission

Submitting a custom MSA is highly recommended: by default, the query MSA is constructed using two iterations of HHblits against uniprot database and filtered to remove very similar sequences. This automatic procedure does not always generate an MSA optimal for a particular case (see note on MSA comparison)

Using the workspace

After submitting the job, you will be redirected to the workspace.

In the workspace, you can:

First, you will see that the job is queued: Job queued in workspace or running: Job running in workspace When it's running, you can monitor the job in detail in the displayed log.

After some time you will see that the job is ready, and you can access the MODexplorer interface by clicking the link as it is shown in the figure: Job ready in workspace If the job crashed: Job crashed in workspace By clicking on '[Display log]' you can see what happened: Job crashed in workspace

Accessing the workspace

There are several ways of accessing the workspace at any time.

Using the MODexplorer interface

You can access the MODexplorer interface by clicking the "access results" link in the workspace: Job running in workspace

You will be redirected to MODexplorer interface. After (optional) interactive tour, you will see this view:

First view of editor with explanations

MODexplorer interface mainly consists of three panels:

Filtering Panel
It allows for filtering the hits by various criteria
Overview Panel
This panel is the core of the MODexplorer interface and contains the list and annotations of the homologs of the query protein that have known structure (i.e. hits).
Detail view Panel
Contains all the information relative to the sequence alignment between the query protein and a hit that the user clicked on. Allows for displaying multiple sequence alignments of families of the query and the given hit.

Filtering panel

The filtering panel is initially collapsed. In order to display it click the arrow next to the word 'Filter' within the header of the panel: Filter_coll Then, you can access all the filtering criteria that are offered by MODexplorer interface. Filter_uncoll

Overview Panel: browse the annotation displayed in MODexplorer interface

In this panel the retrieved homologous proteins of the input query are schematically displayed as BLAST-like bar diagram. The possibility to decide which kind of annotation a user wants to be displayed is offered by the toolbar in the figure: toolbar for annotation

Overview Panel: the hitlist

In the left part of the BLAST-like bar diagram, described before, the PDB ids of the identified homologs are listed. They are in the same row of the corresponding hit bar and their names are in the format <pdb code>_<chain> (ex. 1nq9_A) as displayed in the figure. By clicking the hit name the corresponding PDB entry is opened in a new tab of a Web browser. hit list: pdb codes

Overview Panel: other elements

Within the central panel, over the hit list, there are two elements: a group of buttons allowing the operations described below and a central frame containing residue numbering of the currently highlighted hit residue. hit list: other btns

Detail view Panel

In this panel it is possible to view the selected alignment. In order to see a query-hit alignment in the bottom panel, the hit bar corresponding to the selected hit must be clicked. By default the sequence alignment between query and hit is displayed. The two sequences are separated by the similarity row that is extracted from the HHSearch output.

Default detailed MSA

In the top part of the detail view panel there is a toolbar which contains: Bottom panel toolbar

Exploring ligands and nucleic acids binding sites in Detail view Panel

In "Ligands" and "DNA/RNA" display mode, the Detail view Panel allows exploring individual binding sites in the hits and the sequence similarity of the binding sites to the query.

Upon entering the Detail view for a hit with ligands, the following will be displayed:

First ligand view in detail view

Upon clicking a binding site, the panel will display similarity scores:

First ligand view in detail view

Upon clicking "Display homologs" button, it is possible to evaluate sequence conservation around the binding site across query and hit families.

Ligand view in detail view with homs

Integrated bioinformatics software and databases

HHblits

HHblits is used to build multiple sequence alignment of query and its homologs.

hhfilter

hhfilter is used to filter alignments of query and hit families.

PSIPRED

PSIPRED (version 3.2.1) is run using addss.pl from hhsuite version 2.0.15 package

addss.pl allows to run PSIPRED from custom MSA by calculating the BLAST checkpoint and profile (mtx) file. It performs the following operations (for details please refer to the addss.pl source code):

reformat.pl -v 2 -M first fas a3m target.a3m target.in.a3m
reformat.pl -v 2 -r -noss a3m psi target.in.a3m target.psi
blastpgp -b 1 -j 1 -h 0.001 -d /db/HHsearch/dummydb -i target.sq -B target.psi \ -C target.chk 1> target.blalog 2> target.blalog
makemat -P target
/psipred321/bin/psipred target.mtx /psipred321/data/weights.dat \ /psipred321/data/weights.dat2 /psipred321/data/weights.dat3 \ > target.ss
/psipred321/bin/psipass2 /psipred321/data/weights_p2.dat 1 0.98 1.09 \ target.ss2 target.ss > target.horiz       

The input MSA generated by hhblits and filtered with -diff 100 -cov 5 with hhfilter is used (see above).

ACCpro

ACCpro is used to predict solvent accessibility states of the query and its homologs.

DSSP

DSSP is added based on a local copy of DSSP database updated weekly from DSSP website

If DSSP file is not available in the database, it's calculated on the fly using DSSP CMBI version by Elmar.Krieger@cmbi.kun.nl / November 18,2002, with default parameters.

For display in the editor, DSSP alphabet is converted to three letter alphabet using the following conversion scheme:

G -> H
I -> H
H -> H
E -> E
B -> C
L -> C
T -> C
S -> C

DISOPRED

DISOPRED is used to predict disorder regions in the query

ACCpro

ACCpro is used to predict solvent accessibility states of the target and its homologs.

POPS

POPS is used to calculate solvent accessibility states of the hits.

Modeller

QMEAN

Theseus

Theseus is used to superpose template and model structures. The superposition is performed based on the current query-hit sequence alignment, using only the alignment column range aligned with the query sequence.

SALIGN

SALIGN is used to superpose two selected PDB chains.

Databases

Frequently asked questions

What the server calculates after pressing Submit button?

  1. Creates an MSA of query homolog sequences

    The query MSA is constructed using two iterations of HHblits against uniprot database and filtered to remove very similar sequences.

    This step can be disabled in Advanced options of the submission form.

  2. Generates alignments to protein with known structure (hits) using HHSearch (Söding, 2005) and the PDB database filtered at 70% sequence identity.
  3. Adds CANONICAL SEQRES sequences of hits to the alignments of the hits

    The sequences are aligned to corresponding original sequences already in the query-hit alignment using muscle program with the following parameters:

    -maxiters 1
    -diags
    -sv
    -distance1 kbit20_3
    -gapopen -100

    According to our tests, these parameters ensure fast accurate alignment of two sequences that are nearly identical but can contain missing regions relatively to each other.

  4. Prepares MSAs of representative homologs of the query and the hits

    Filtered MSAs of hit homologs are retrieved from HHSearch database of alignments.

  5. For every HHSearch hit, retrieves and aligns related PDB chains based on information about omitted PDB chains contained in HHSearch database and based on clusters of similar PDB chains from MMBD MMDB Non-redundant PDB chain set.
  6. Adds ATOM sequences

    ATOM sequences are derived from ATOM record of PDB files. Some non-standard amino acids are changed to canonical ones if possible. The sequences are aligned to corresponding SEQRES sequences using muscle as above.

  7. Predicts secondary structure for representative homologs

    Secondary structure is predicted using PSIPRED

  8. Predicts solvent accessibility for representative homologs

    Solvent accessibility is predicted using ACCpro

  9. Adds secondary structure for hits

    Secondary structure is calculated with DSSP.

  10. Predicts disorder regions for the query

    Disorder is predicted with DISOPRED.

How to get Modeller key?

To obtain Modeller key submit Modeller registration form online. MODELLER is available free of charge to academic non-profit institutions.

Which is the best browser for MODexplorer?

MODexplorer works well in all major browsers.

Exact performance depends, however, on the particular operating system / web browser combination.

Browser/OS Linux MacOS Windows
Google Chrome v. 16.x
Safari v. 5.1.x
Firefox 8.x - 9.x
Firefox 3.x - 7.x
Opera
Internet Explorer 6.x - 9.x

Data for: January 2012

Legend
Best combination
It works and performs acceptably
It works, but SLOW performances
It works, but VERY SLOW performances
It doesn't work at all
Browser not available for that OS

How long jobs are kept on the server?

Jobs are kept for 90 days from the last access

How is privacy of my job protected?

Your jobs get unique id numbers composed of submission date and random string composed of numbers and letters (e.g. 2011-10-13/kP9aS_ru1z). Your jobs can be accessed from the web browser you run it only as long as the browser data is not cleared. From all other browsers in the world, it can be accessed only using that difficult-to-guess job id. If you name your job "p53" there is still no way users can access your job by this common name - the job id described above is always required.

If you run or access your job from a public web browser we recommend:

  1. Provide e-mail address in a submission form
  2. After finishing the work on your job, clear cache, cookies, and history of your browser OR/AND:
  3. Delete the job in Your recent jobs section.

    The job will be deleted only from the web browser data, you can still access the job using a link in the e-mail

How to cite MODexplorer?

Manuscript in preparation

Troubleshooting

Slow interface

MODexplorer editor can get slow when:

We recommend the following ways of solving performance problem:

3D visualization (Jmol) is not working

Platform specific notes:


Please send feedback to: support@modorama.org


If you use this web server, please cite the following reference:

  • Kosinski, J., Barbato, A., Tramontano, A. MODexplorer: an integrated tool for exploring protein sequence, structure and function relationships (2013) Bioinformatics, 29(7):953-4 [article]