Protein Association Network Analysis Using STRING (Part 1)

Protein Association Network Analysis Using STRING (Part 1)


STRING is a database of
functional associations. So let’s search. Go to Google, and let’s
search for STRING. So this is how I search
for STRING. Type string, and then make sure
you also for protein. So there is your first hit,
STRING– functional association networks. Click on that. That takes us to the
STRING database. And like I mentioned, STRING
is a database of known and predicted protein-protein
interactions. One of the great, nice things
about STRING is you work on any organism– most of the organisms– almost every single organism
that has a complete genome sequence available has
interactional association information in STRING. There are many databases
out there. Some of them have data only
for human beings. Some of them have data only
for the mammalians. Some of them have plants. But the nice thing about STRING
is it has associations for every single known genome,
be it a bacteria, be it a protozoa, be it an animal– mouse or human. So you always have something
here for anybody that’s working out there. All right, so you can search
STRING database for a single gene or protein. You can search it by name,
or you can use a protein sequence to search. But what we’re going to do
today– what I’m going to show you today– is using
multiple names. So the click on this
option here. So that gives you an input form
where you can list you genes of interest. Let me grab a list of gene. I have 44 genes that were
differentially expressed in my current experiment. So I’m going grab that, Select
All, Copy, and then I’m going to Paste it in STRING. So there is my 44 genes of
interest, and as I mentioned, I know that these are
co-expressed in my current experiment. So I already know that there
could be some kind of association in them, among those
genes, but I don’t know what kind of association
is there. So that’s what I’m
looking up here. And again, specific organism– I know it’s human,
so I pick a type. And you can pick one
that’s appropriate. So I click select the
humans there. And then I hit the button, Go. As soon as I hit the button, Go,
what’s going to happen is it looks up for every single
gene that I input, and it checks with that. And it gives you an option
to either include it or exclude it. So I check everything is fine,
so I’m going to continue. I click this button, Continue. It then takes me to the
actual results page. This is a page that’s
going to have– if there is any association
among these genes from 44 genes of my interest, that’s
going to show up here. But as you see, there are 44
genes from my list, but there are three associations,
three lines. Every dot or circle represents
a protein, and the lines represent the association. There are only three lines
here, three associations. That’s not– maybe interesting, but there
may be something more here. So there are a couple of ways
you can play with this. STRING is for you to
do basic analysis. This is not an extensive
network analysis tool. It’s a database, interaction
database, but it also has basic network analysis
capabilities. I’m just going to show you
a couple of them here. If you scroll down the page,
you get options to choose parameters here. And like you see, these
are the matters– gene neighborhood, gene fusion,
co-occurrence, co-expression, experimental data, databases,
textmining. These are the sources for all
the associations that STRING shows up. And you’ve have an option
to choose a level of confidence here. Usually, by default, it’s
the medium confidence. But since with the medium
confidence, we didn’t find anything interesting, so I
choose low confidence here. And then what you want to do
is update the parameters. See this little button here. So click that button,
Update Parameters. When you do that, it’s going to
show associations that were not previously shown because
of the confidence level. So now you see, even though
we reduced a little the threshold, previously we saw associations that are separate. But they are now together. So you see this nice little
network here. So while you could say that
the level of confidence is less here, but as I mentioned
earlier, already I know that this is the list of genes that
come from a differentially expressed experiment of from a
microarray experiment where I know that they are all
correlated or co-expressed. So I’m already suspecting
some associations. Now, this gives me the
potential of possible association among them. There are a couple of things
that I wanted to quickly show you here. If you click on any on this
line, it’s going to show you what the score is, and it will
also show you what that potential interactions is. For example, SNX3 is associated
with LDA. And you can see that says core,
and it comes from the co-expression. If you’re interested in
exploring more about where the co-expression is, you can
click this Show button. It will show you the evidence
for that co-expression. I’m going to close this here. And here, you see two lines
because the evidence comes from two different resources. And let me also click
another one. Let me click this interaction
association. And this one, as you see,
it is co-mentioned in permanent abstracts. And there this code here. And if you wanted to see the
exact abstract, exactly how they are co-mentioned in the
paper, you can click this and that will show you. Let’s try that. I’m going to click this.

One thought on “Protein Association Network Analysis Using STRING (Part 1)”

  1. Is there a software where you can make the same analysis but considering the genes/protein fold changes? Thanks in advance

Leave a Reply

Your email address will not be published. Required fields are marked *