Can anyone help me to perform protein-protein interaction network analysis using cytoscape for non-model plants. I am using stringapp in cytoscape, but my plants are not there in the organism list of stringaap. Then what to select?
I want to analyze DE proteins form whole proteome data, for thier interactions. I have Uniprot IDs of identified proteins.
Dear Shweta Jha
you can use a model organism as the basis for your network.
Example: Arabidopsis thaliana
Or another plant that is more studied than yours and closest in the tree of life.
Best,
Leite
Hi, If your plant is not in the species list, it will be hard to construct your protein-protein network. Did you check if protein-protein interaction information are available in another database? BioGRID database gives only PPI and the type of evidence. String database can be messy with Text mining interactions.
If there is no PPI information, the solution is to find orthologs of your protein of interest in plant models as A thaliana. It can be informative too.
Hervé
Thanks for your good suggestion. I have UNIPROT ID of proteins from different organisms, identified from proteomics data. How can I find orthologs of those proteins in Arabidopsis thaliana? Should I just download the protein sequences from UNIPROT and BLAST in NCBI or TAIR? or is there any other strategy?
Hi,
Do you have any proteomic data from your non-model plant? If data are available, you can first homogenize mass spectroscopy data.
For each protein from MS, blast against your non-model plant protein database and identify the best candidate (High score, high identity, best p-value). This step will homogenize your data, and you will have data only annotated in your non-model plant. If proteome of your non-model plant is not available, you can use proteome from the same family to homogenized.
After that, you can use Orthofinder:
http://www.stevekellylab.com/software/orthofinder
This methodology is really powerful and avoids bias from length sequences that you can face with classical Reciprocal Best Hit.
When you have orthologs you can use it for construct and analyze the PPi.
We used a similar strategy for papers in my PhD group:
Article Toward the understanding of the role of CDC48, a major compo...
Article Functional characterization of the chaperon-like protein Cdc...
I hope it can be helpful.
Hervé
Thanks a lot Dr. Hervé Begue for your valuable suggestions. I do not have proteomic data available for my non-model plants, so I'll use proteome form the nearest plant.
Can Orthofinder be used on Windows? I am not used to work on Linux.
Here find some links for running Orthofinder: http://www.stevekellylab.com/software/orthofinder https://github.com/davidemms/OrthoFinder/blob/master/OrthoFinder-manual.pdf
Unfortunately, it is written in the manual provided in the second link that Orthofinder runs with Linux and Mac. As you will read, the use of Orthofinder is effortless. You need only one code line, one file with different proteome in Fasta file (one Fasta file by proteome) and that's it. You also need different dependencies, but all are written in the manual.
Little advice, if you will perform this analysis, you should use Diamond as alignment-tool instead of BLAST. Diamond is incredibly faster!
I hope you will enjoy this tool.
Hervé
Hi Hervé Begue ! What do you usually do when you have multiple orthologs for one gene? Do you choose just one of them for downstream analysis, such as GO term enrichment/STRING?
- Badges
- Science topic
- Similar topics
- Biological Science
- Biota
- Eukaryota
- Fungi