Max-Planck-Institut für Informatik
max planck institut
mpii logo Minerva of the Max Planck Society


The core computation pipeline can separated in three major steps.
  1. The pipeline collects the amino acid sequences of the proteins corresponding to the list of given identifiers.
  2. The pipeline searches for resolved 3D structures of the respective proteins. In this step, the structures of homologous proteins are also considered. A BLAST search of the up-to-date PDB instance is performed with the E-value threshold of 10-10.
  3. The spatial location of the amino acid residues affected by the mutations is analysed: The distances to all other protein and nucleic acid (DNA and RNA) chains and hetero atom groups co-crystallyzed with each of the identified homologs. This is called structural annotation. The common buffer components are ignored. The distances, the similarity between the query and the structurally resolved proteins, and the quality of the corresponding structure are combined into an Interaction Score that describes how likely the nsSNP is to disrupt an interaction.
Pipeline Chart

All the distances, contacts and scores are stored in a MySQL database. The database connects the produced annotations with the input and all informations gathered on the way in order to produce outputs generated by different additional analysis options.


Alexander Gress, Vasily Ramensky, Joachim Büch, Andreas Keller, and Olga V. Kalinina
StructMAn: annotation of single-nucleotide polymorphisms in the structural context
Nucleic Acids Research, first published online May 5, 2016


This is a simple tutorial going step-by-step through the most basic functions of StructMAn.

1. Uploading an input file

The simplest input file for StructMAn contains only information on mutations that need to be structurally studied. Press on the green arrow button as shown in the picture below a file from your system. For bigger files, the upload may take some time, especially for bigger vcf-files. The format of the uploaded file is checked and you will get an immediate response, if the file format was not correct. Mutation lists can be uploaded in the standard VCF format or a custom SMLF format.
Upload Screenshot

2. Choosing output options

The results are presented in a form of a table sorted either by the mutation interaction score or the protein score. The order can be switched to protein scores by selecting the "Annotations" and "Proteinsort" options. Selecting "Annotations" alone will result in ordering by interaction score.
Output Options Screenshot

3. Expert Options

Press the "Show options" button to expand the expert options window. Expert options allow to control the thresholds set for similarity search, structure selection and scoring.
Expert Options Screenshot
For this basic tutorial choose "default". If you do not provide any expert options, the server will also use these default values.
Default Options Screenshot

4. Submit your query

Press the "Submit" button.
Submit Button Screenshot

5. Wait for the results

After pressing the submit button, the server puts your job submission into a working queue and redirects you to your personal results page. You can bookmark this page and come back to it later. The page shows you how many jobs are in the queue before your job.
Running Job Screenshot

6. Viewing and downloading results

The results come as an interactive table, but are also downloadable via the "Download all results" button. In the table, each row corresponds to a single mutation that can be mapped to a 3D structure of the same protein or a homolog. The results are ordered using the interaction score or the protein score.
Download Results Screenshot

7. Changing query options

If you want to change query options or the output mode, you have to resubmit your job, but it will run virtually instantaneous using the produced data of the first job.

8. 3D-visualization of the mutations

Press on any "JSmol"-Button to open a new tab in browser showing the 3D-visualization of the mutation from the corresponding table row. This may take some time to load depending on the size of the displayed structure. The initially displayed snapshot is centered around the residue, which corresponds to the substituted one.
Visualization Screenshot