Software

Our ambition is to develop better methods for prediction of protein structure and function.
The most successful methods are provided as web-based servers.

Git repositories

• Main directories
  • Github https://github.com/ElofssonLab
  • BitBucket https://bitbucket.org/ElofssonLab
• PconsC3
• ProQ3 https://bitbucket.org/ElofssonLab/proq3
• PconsC4 https://github.com/ElofssonLab/PconsC4.git and the model
  
   

  2) Get the trained model (it would be good if we could add it to git, but our git lfs is full): 

  cd PconsC4

  wget https://www.dropbox.com/s/kgyz31he9q6fkoi/pconsc4_unet.python3.h5 -O src/model/pconsc4_unet.python3.h5
   
  3) Run PconsC4 on an aligment:

  python src/run_pconsc4.py <path_to_alignment.a3m> [-o <path_to_output_file> -t <number_of_threads> -s <min_sequence_separation>]

  Compiling the cython source code is now done automatically when running “src/run_pconsc4.py” for the first time. So no need to do it manually with the provided setup.py script anymore.

Web-services developed through EMBRACE

• Pfrag: a program for homology modelling
• SHRIMP: a program for Hidden Markov Model-based homology scoring

Structure Predictions

• PconsD: ultra rapid, accurate model quality assessment for protein structure prediction.
• PconsC: combination of direct information methods and alignments improves contact prediction.
• PconsC2: Improved contact predictions using the recognition of protein like contact patterns
• Pcons.net: The Pcons meta and consensus server. Now it also builds very nice homology models in less than a minute .
• ProQ: a predictor of protein quality.
• TOBMODEL: automatic modelling of beta-barrel membrane proteins.

Membrane Protein Topology Prediction Methods

• TOPCONS: A new consensus method to predict the topology of membrane proteins. Now features signal protein prediction and uses a very fast method to create profiles that enables prediction of entire genomes.
• TOPCONS-single: a consensus method for topology prediction of alpha-helical membrane proteins, but without the need for BLAST runs. It is based on several tools that accept a single protein sequence as input. Ideal for a quick scan of entire genomes.
• SCAMPI: topology predictor for beta-barrel membrane proteins. Ideal for a quick scan of very large datasets.
• BOCTOPUS: topology predictor for beta-barrel membrane proteins
• OCTOPUS & SPOCTOPUS: prediction of membrane protein topology and signal peptides
• ZPRED: prediction of TransMembrane Regions and distance to the membrane center (available through the TOPCONS webserver)
• DAS: prediction of TransMembrane Regions.

Subcellular Localization Prediction Methods

• SubCons: a new ensemble method for improved human subcellular localization predictions.