GraPES currently houses the MaGS (Membraneless organelle and Granule Score) and MaGSeq
tools that predict the propensity of proteins to colocalize into liquid-liquid phase separated biological
condensates by evaluating a protein and providing the user with a z-score as compared to the proteomic
distribution of human or Saccharomyces Cerevisiae cell lines. These tools are parameterized using the
same workflow but differ in their implementation:
- MaGS uses protein-specific, experimentally measured data in its evaluation and will provide the
user with the most accurate assessment. However, this method is more limited as it requires all
experimental data to be available. Currently, users can search a database of pre-computed
scores over the human and yeast proteomes.
- MaGSeq uses only the primary sequence of a protein in its evaluation and, while not as accurate
as MaGS, provides the user flexibility to test novel protein sequences or to make predictions of
proteins of other eukaryotic organisms.
For MaGS (currently on the home page) simply query the protein of interest by UniProtID or gene name
and the webtool will pull up any entries in the database that are available.
For MaGSeq, navigate to the MaGSeq tab at the top of the web page then you should be able to enter
the following information: FASTA sequence(s), Model (Human/Yeast) and E-mail (optional). When you
have entered the information, click the Submit button.
After your jobs have run, you will be able to access a variety of output data. Most importantly, is the Z-
Score, as this represents the propensity for localization into biological condensates. While there is no
strict cutoff, we would recommend a score of over 1.5 as an initial threshold keeping in mind that phase
separation in a biological environment is complex and nuanced. A number of other features are included
in the output as well, and can be used to help inform which features are helping to drive the prediction.
Each of the features and the z-score are represented graphically:
Where the y-axis is the Density (percent) of proteins with a given score (the x-axis). The background
distribution of that feature in the proteome is shown in gray. Then a number of known foci markers are
shown in addition to the queried protein. Here, for instance, the PScore is shown with the markers
DCP2, PAB1, and G3BP1 along with the queried protein. The protein of interest has a much higher
PScore, which is a measure of increased pi-pi interactions, than the markers so the user can conclude
that this property is likely to aid the protein in forming multivalent interactions and to form
condensates.
In addition to the graphical outputs a plain text file can be downloaded containing data for that protein.
This file is formatted as:
{
"accession": Unitprot Accession,
"species": Species,
"score": MaGS Z-Score,
"abundance": Abundance,
"camsol": Camsol,
"annotatedPhosphorylationSites": Annotated Phosphorylation Sites,
"disorder": % Disorder,
"compositionD": % Composition of D,
"compositionE": % Composition of E,
"compositionL": % Composition of L,
"compositionG": % Composition of G,
"rna": RNA-binding protein (true/false),
"sg": Known SG protein (true/false),
"pScore": Pi-Pi PScore
}
If an email address is provided (providing an email address is optional), a notification email will be sent
once a job is processed. Notification a link to the results page for the job.
At the bottom of the Home and MaGSeq tabs there is a table which provides the user information on
the queuing system and the state of their submitted jobs. Once your jobs have been finished your
results can also be accessed through this table.