Rational Design

For DMR lab.

This program (design) generates a list of candidate peptide sequences for inhibition of a desired protein-protein interaction, using the rational design method developed in the Mochly-Rosen lab.

Inputs: Two protein names and their Uniprot accession IDs OR Two .fasta protein sequence files from UniProt.

Output: A .pdf report containing homology, alignment, and heatmap information for the specified number of candidate peptides.

Usage:

You can input protein names and their Uniprot accessions:

$ design -p1 protein1_name protein1_accession -p2 protein2_name protein2_accession -l peptide_length -o name_of_output_report.pdf
# for example:
$ design -p1 PDK Q15118 -p2 DPKC Q05655 -l 5 -o pdk_dpkc_report.pdf

Or, you can input two protein .fasta files:

$ design -fasta1 protein1.fasta -fasta2 protein2.fasta -l peptide_length -o name_of_output_report.pdf
# for example:
$ design -fasta1 test/pdk.fasta -fasta2 test/dpkc.fasta -l 5 -o pdk_dpkc_report.pdf

You can specify a range of peptide lengths, for example -l 5-10. If you have a PDB ID for one or both proteins, you can add them with the flags -pdb1 or -pdb2

$ design -p1 Drp1 O00429 -p2 Mff Q9GZY8 -l 5-10 -o drp1_mff.pdf -pdb1 4BEJ

General Workflow

1. Get a list of all possible peptide alignments with homology scores
   • Try a few different amino acid lengths (ex. 5-10), specified by user
   • Keep only peptide pairs with homology score greater than or equal to 2 stdev above the mean
   • Combine overlapping peptides into one entry
2. For each homologous peptide pair:
   • Calculate the conservation of each peptide across the list of organisms
   • Filter out the ones that are poorly conserved? This is currently not done, but can be done by changing filter_by_cons=True in design.py
3. For each homologous peptide pair:
   • Blast to get other proteins that contain that peptide (exclude putative and fragment proteins)
   • Calculate the conservation of that peptide in that protein across the list of organisms
   • Make a heatmap showing the 10 proteins with the most conserved homologous peptide
4. Optional: Structural information from input PDB ID
   • Calculate and display relative accessible surface area
   • Display secondary structure
5. The report includes:
   • Summary Page
     • Table of peptides
       • Homology score
       • Peptide sequence
       • Peptide position
       • Peptide conservation
   • Peptide pages
     • Summary of similar (shorter) candidate peptides
     • Structural information, if PDB ID provided
     • Alignment of each peptide
     • Heatmap
     • Table of protein descriptions

Installation

This program is meant to be run in a virtual machine (VM) with all the prerequisites already installed. This allows anyone on any operating system to run the program without complicated and OS-specific installation issues. To start the VM, you will need Vagrant and VirtualBox.

1. Download and install Vagrant here: https://www.vagrantup.com/downloads.html

2. Download and install VirtualBox here: https://www.virtualbox.org/wiki/Downloads

3. Download this repository (Releases > v1.0 > Download Source code (zip)) or clone it:

$ git clone https://github.com/annadcunningham/rational-design.git

In your Terminal, initialize the VM. This may take a while (10-20 minutes).

$ cd rational-design/vagrant
$ vagrant up

4. Enter the VM and navigate to the design directory:

$ vagrant ssh
$ cd /design/

If you are using .fasta input files, you should save them somewhere in this design directory so you can access them within the VM.

5. Now you can run your proteins of interest (see Usage section above).

6. When you're all done, exit the VM:

$ exit

You can always re-enter it using vagrant ssh (always do these commands from the rational-design/vagrant directory). You can leave the VM running; it will take up 2GB of RAM. Or you can delete it:

$ vagrant destroy

To check whether you have the VM running:

$ vagrant status