Transcript C-metabolism-tour - BioBIKE Portal

Combining the microarray and
metabolic capabilities of BioBIKE
Case Study
How is carbon metabolism affected by starving
the cyanobacterium Anabaena for nitrogen?*
This demonstration is best viewed as a slide show,
enabling you to simulate a session and make
changes in cursor position more obvious.
To do this, click Slide Show on the top tool bar, then View show.
*Problem taken from Xu X, Elhai J, Wolk CP (2007). Transcriptional and developmental responses
by Anabaena to deprivation of fixed nitrogen. In: The Cyanobacteria: Molecular Biology, Genetics
and Evolution (Herrero A, Flores E, eds). Horizon Scientific Press. pp.387-426.
Nitrogen deprivation
differentiation in Anabaena
+N
All cells in same state
+N
Autofluorescence
(photosynthetic apparatus)
The cyanobacterium Anabaena grows as
multicellular filaments, all cells in the same state
when nitrogen is plentiful.
Anabaena is photosynthetic, so it supplies its
own carbon from CO2.
Nitrogen deprivation
differentiation in Anabaena
+N
+N
Autofluorescence
(photosynthetic apparatus)
All cells in same state
-N
N2-fixing
heterocysts
When starved for nitrogen,
Anabaena differentiates
specialized cells, heterocysts,
that can utilize atmospheric
N2 gas.
Metabolic correlates to N-deprivation
Photosynthetic
apparatus
Amino
acids
proteins
N2-fixing
heterocysts
When the supply of fixed nitrogen (e.g. ammonia)
dries up, Anabaena faces the necessity of making
N2-fixing heterocysts fast. They get the amino
acids to make the new protein needed by
heterocysts by cannibalizing the biggest source of
proteins in the cell – the photosynthetic apparatus.
Metabolic correlates to N-deprivation
Photosynthetic
apparatus
Amino
acids
proteins
glycolipid
Sugar
Energy
But new cells are not just proteins. The surface is
covered with special glycolipid, made in part
from sugar. Sugar is also the source of energy
that drives the process of differentiation.
If photosynthetic capacity is being depleted, then
where does this sugar come from?
N2-fixing
heterocysts
Metabolic correlates to N-deprivation
What enzymes of carbon metabolism are affected by N-starvation?
RNA
-N +N
We can address this question by examining
which genes involved in carbon metabolism are
turned on or turned off by N-starvation.
Microarrays can be used to get a snapshot of
gene expression under different conditions.
Microarray
Courtesy of Inst. für Hormon-und Fortpflanzungsforschung, Universität Hamburg
Metabolic correlates to N-deprivation
What enzymes of carbon metabolism are affected by N-starvation?
Two broad sets of enzymes are of interest:
- The enzymes that catalyze the dark reactions of photosynthesis
- Those that breakdown glycogen and metabolize glucose derivatives.
Metabolic correlates to N-deprivation
What enzymes of carbon metabolism are affected by N-starvation?
Glycogen
metabolism
Pentose
Phosphate
Pathway
Cyanobacteria use primarily the reactions of the Pentose Phosphate
Pathway to break down glucose derivatives. They use carbon fixation
reactions to build glucose. These sets overlap a great deal.
Carbon
fixation
Metabolic correlates to N-deprivation
PLAN
- Define genes involved in pentose phosphate pathway
- Find from microarray how the levels of expression
of these genes are affected by N-deprivation
- Sort the results, from lowest expression to highest
- Display the results, along with a description of the genes,
in an intelligible way
Go to the BioBIKE Portal
biobike.csbc.vcu.edu
Go to a public site
Your name (no spaces)
Enter a log in name
Click New Login
Function palette
Workspace
The BioBIKE environment is divided
into three areas as shown. You'll bring
functions down from the function
palette to the workspace, execute
them, and note the results in the
results window
Results window
Two very important buttons on the
function palette:
HELP!
On-line help (general)
PROBLEM
Something went wrong?
Tell us!
Two very important buttons in the
workspace:
Undo (return to workspace
before last action)
Redo (Get back the
workspace you undid)
Pentose
Phosphate
Pathway
What now?
The first step in exploring the expression of the genes of carbon
metabolism is to define each set. In this tour, we'll focus only on
the genes of the pentose phosphate pathway, but you will be able
to imagine how the analysis could be extended.
Mouse over the DEFINITION button
… and click DEFINE, to bring down a
DEFINE box.
DEFINE box? Before going on, let's take
a break to look at what boxes mean.
General Syntax of BioBIKE
Function-name
Argument
(object)
Keyword
object
The basic unit of BioBIKE is the
function box. It consists of the
name of a function, perhaps one or
more required arguments, and
optional keywords and flags.
A function may be thought of as a
black box: you feed it information,
it produces a result.
Flag
General Syntax of BioBIKE
Function-name
Argument
(object)
Keyword
object
Flag
Function boxes contain the
following elements:
• Function-name (e.g. SEQUENCE-OF or LENGTH-OF)
• Argument: Required, acted on by the function
• Keyword clause: Optional, more information
• Flag: Optional, more (yes/no) information
General Syntax of BioBIKE
Function-name
Argument
(object)
Keyword
object
Flag
… and icons to help you work with
functions:
•
Option icon: Brings up a menu of keywords and flags
•
Action icon: Brings up a menu enabling you to execute
a function, copy and paste, information, get help, etc
Clear/Delete icon: Removes information you entered
or removes box entirely
•
General Syntax of BioBIKE
Function-name
Argument
(object)
Keyword
object
Flag
… and icons to help you work with
functions:
•
Option icon: Brings up a menu of keywords and flags
•
Action icon: Brings up a menu enabling you to execute
a function, copy and paste, information, get help, etc
Clear/Delete icon: Removes information you entered
or removes box entirely
•
And now back to our show…
The DEFINE function takes two arguments:
the name of the variable to be defined and its
value. We want first to define the set of genes
encoding the enzymes of the pentose
phosphate pathway.
Click the variable argument box.
That opens the box for input. Notice that the
box is now white and selected (surrounded
by red dots).
Type the name you want to give the set. The
name must not contain spaces, but symbols
are OK. I chose pentose-phosphate-genes.
The box is still open, and the function will
not be able to be executed until it is closed.
To close it, you could press Enter on your
keyboard. Alternatively (and better in this
case), press the Tab key to close the box and
move on to the next box.
Now the value argument box is open for
input.
We'd like to put in that box the genes of
Anabaena involved in the pentose phosphate
pathway. To get a function that will provide
those genes, mouse over the Genome
button…
… and click GENES-IN-PATHWAY/S.
This function calls for two arguments: the
pathway we're interested in and the organism
(or a replicon from the organism). The
organism is clear, so let's do that first.
Click on the organism-or-replicon argument
box. Then, once it's open, mouse over the
DATA button to get the organism, Anabaena.
Anabaena is a nitrogen-fixing
cyanobacterium, so mouse down
to that category
… and across Anabaena PCC 7120 to…
Hmmm. Looking ahead, I realize that the
microarray I'll be using considers only
chromosomally encoded genes. It will
simplify matters if I choose the
chromosome of Anabaena. Click that.
Now we need to supply the pathway.
Eventually it will be possible to choose
the pathway from a hierarchical menu in
the same way you chose the organism.
For now, the task is more complicated.
We need to find the pentose phosphate
pathway from within a list of pathways
and then copy its identification number.
To find it, we'll display the list.
Go to to the INPUT-OUTPUT button
to find the DISPLAY-LIST function.
Click DISPLAY-LIST to bring down
the function.
DISPLAY-LIST formats lists of
information to make them easier to read.
We're going to use it to examine lists of
pathways to help us find the one we
want.
Click the list argument box, then mouse
over the DATA button to get to lists of
pathways.
Mouse over kegg-pathways, and click
*CARBOHYDRATE-METABOLISM*,
the most likely category in which to find
the pentose phosphate pathway.
Carbohydrate metabolism consists of
many pathways. We want to display each
one, with formatting, so mouse over the
EACH prefix option and select it.
The function is now complete and ready to
be executed. Mouse over to the action icon
of DISPLAY-LIST…
…and click Execute.
A popup window appears, displaying the list of pathways defined by
KEGG as related to carbohydrate metabolism. The pentose phosphate
pathway has an identification number "00030". That's what we came for.
Important displays can be downloaded to your computer through the usual
browser controls. This time just X out of the box.
With the pathway identification number
in hand, click the pathway argument
box…
… enter "00030", and press the ENTER
key. Once you've done that, all the boxes
will be filled in and closed. The function
is now ready to be executed.
Mouse over the Action Icon of the
DEFINE function…
…and click Execute.
Note that a new VARIABLES button
appears. We'll use it in a moment. At the
same time, a list of genes appears in the
Result Window.
What is that list of genes? It's supposed
to be the genes in the chromosome of
Anabaena that encode enzymes of the
pentose phosphate pathway. Is that true?
To check, go to the GENES-PROTEINS
button,…
… and click DESCRIPTION-OF.
We can use this function to describe the set
of genes we just defined. If we recognize
within the descriptions names of enzymes
of the pentose phosphate pathway, all is
OK.
The function is asking for something to
describe. Click on the entity argument box.
The set we just defined has become part of
our language, accessible through the
Variables button.
Mouse over that button…
… to find pentose-phosphate-genes.
Click on it to bring it down to the argument
box.
You can execute this function as it stands,
but the result will be disappointing (try it!).
To make the result more intelligible, mouse
over the Option Icon…
… and click DISPLAY. Now the results
will not only be presented in the Result
Window, they will also be displayed in an
intelligible form.
You might think that now the function is
ready, but try it and you'll get a description
of the list (it's a list) when what you want is
a description of each element of the list. So
as before, you need to specify the EACH
prefix option.
Finally, execute the function.
Pentose
Phosphate
Pathway
Glucose-6-phosphate dehydrogenase (G6PDH), transaldolase... They all seem
to be there. You've done the right thing, checking the result. Now kill the box.
PLAN
We've defined the genes, now we
define their expression levels.
Mouse over the DEFINITION
button.
Define genes involved in pentose phosphate
 - pathway
- Find from microarray how the levels of
expression of these genes are affected by
N-deprivation
… and click DEFINE.
Define genes involved in pentose phosphate
 - pathway
- Find from microarray how the levels of
expression of these genes are affected by
N-deprivation
As before, we put in the variable
box a memorable name for the set.
Click the variable box.
… and type in a logical name
(I chose expression-levels).
Then press TAB.
Now to get the microarray
data. Mouse over the
GENOME button to reach the
microarray functions.
…and click INSIDEMICROARRAY to gain
access to the expression levels
of specific genes.
INSIDE-MICROARRAY
needs to know the name of the
microarray experiment, the
names of the genes your
interested in, and the
experimental condition.
There are other ways, but
suppose we already know of a
pertinent experiment…
To use the data of Ehira
and Ohmori, click the expt
argument box…
…and go to the DATA
button, where microarray
data can be found.
Navigate to Ehira and
Ohmori's experiment,
and click it.
Experiment in place, where to find
the genes? We've already defined
the set we're interested in, so open
the argument box and mouse over
the VARIABLES button.
… and select our set.
Now for the experimental condition.
For this we need to understand how
the experiment is described.
We've accessed DESCRIPTION-OF
before, but another way of finding it
is alphabetically. Mouse over the
ALL button.
Find and click DESCRIPTION-OF.
We could get the experiment the same
way as before, but it may be easier to
copy and paste what we already have.
Mouse over the Action Icon of the
experiment.
... and click Copy. The contents of the
box is placed in BioBIKE's clipboard
(different from the clipboard of your
operating system)
To paste what's on the BioBIKE
clipboard, mouse over the Action Icon
of the target box.
… and click Paste.
If we executed the function now, we’d
learn that ehira_2006 is an experiment,
not a profound insight. To learn more
than this, mouse over to the Option
Icon,…
…and click FULL to open up a frame
containing all the information pertaining
to the experiment.
Click Execute, as usual.
We get a table of contents for the
experiment, which we can expand,
if we like. But the information
we’re looking for is already here.
Condition 2 is the microarray that
measures expression 8 hr after
nitrogen deprivation.
X out of the box.
Now we can fill in the last argument
box,…
…with the condition number, 2.
(Remember to close the open
argument box by pressing Return).
We still have to specify what
information we want out of the
microarray.
Mouse over the Options Icon.
We can choose as much information as
we like, but the log expression ratio is
enough. That number represents the
ratio of expression at 8 hr of starvation
divided by the control experiencing no
starvation, represented as a log so that
increased levels of expression are
positive numbers and decreased levels
negative.
Now execute
The Result Window provides the desired answer –
gene + log2-expression-level – good for further
computation but not good for human consumption.
We used DISPLAY-LIST before to format such
results. We can make use of that function box now.
First, clear the earlier argument.
(BioBIKE moved the now smaller function up a
line to conserve space).
Select the empty list argument box and fill it with
the variable you just defined, accessible through the
VARIABLES button.
Select the new variable.
(BioBIKE moved the now larger function back to
where it was before. This is a bit jarring, but space
is valuable!).
Mouse over the Action Icon of DISPLAY-LIST to
execute.
Click Execute…
Well, we got what we asked for: genes and their
expression levels, but somehow it doesn’t burst
with meaning.
It would be more useful if there were some
description of the genes and if the expression levels
were sorted.
X out of this box.
First the description. My strategy is to define a set
of descriptions of all the genes and then fold it into
the set of expression levels.
Mouse over the DEFINITION button…
…and click DEFINE.
Name the variable something comprehensible, and
tab to go to the value argument box.
We’ve already used the DESCRIPTION-OF
function. Since the argument box is already
selected, all I need to do is to cut out the
DESCRIPTION-OF box and insert it into the
open box.
To do this, mouse over the box’s Action Icon…
…and click Cut/Insert
Since we don’t want to look at the descriptions
(we’ve already seen them) but rather work with
them, we can get rid of the DISPLAY flag.
All good. Now execute it, by mousing to the
Action Icon…
…and clicking Execute.
We now have a list of gene descriptions to
interleave with the previously existing list of
expression levels.
To do the interleaving, mouse over the LISTTABLES button…
… and navigate to INTERLEAVE.
INTERLEAVE asks us to provide two lists. The
first…
…is the expression levels we defined earlier.
The second…
…is the list of descriptions.
If we execute it as it stands (try it), you’ll be
reminded that the expression-levels consist of
two values, which will be separated from the
descriptions by parentheses. To avoid this fine
distinction, SIMPLIFY the list at each round of
interleaving.
You can execute this function to make sure it
does what it is supposed to do (always a good
idea), but in the end, we want to display the list
as we did before.
To do this, we can delete the old argument to
DISPLAY-LIST…
…and move the INTERLEAVE box into the
hole of DISPLAY-LIST.
Click somewhere in the INTERLEAVE box,
holding it for a half a second…
…and drag it to the list argument hole.
…then release it.
Finally, mouse over the Action Icon…
…and execute.
Better!
But not sorted.
X out of the box
To sort the list, we want to somehow put the
INTERLEAVE box inside of a SORT function.
Not difficult. Go to the Action Icon of
INTERLEAVE.
…and click Surround with
This selects the box (notice the red
dotted outline).
Now mouse to the LIST-TABLES
menu and click on SORT.
SORT is now wrapped around the
INTERLEAVE box, so the list
will now be sorted.
But sorted how?
We could sort by any of the three
columns of the list. We want the
second.
Mouse over the Options Icon…
…and click BY-POSITION…
…select the value box…
…and enter 2, for the 2nd column.
Now execute.
Much better! We can readily
identify those genes of the
pentose phosphate pathway
that are up-regulated by
nitrogen-starvation.
What about the genes of
carbon fixation?
The steps are the same as
before… we’ll fast-forward
to the result.
Metabolic correlates to N-deprivation
What enzymes of carbon metabolism are affected by N-starvation?
Glycogen
metabolism
Pentose
Phosphate
Pathway
Putting the expression results for the three pathways together...
Carbon
fixation
Metabolic correlates to N-deprivation
What enzymes of carbon metabolism are affected by N-starvation?
Glycogen
metabolism
Pentose
Phosphate
Pathway
We can see that the portion of the pentose phosphate pathway shared
with carbon fixation is repressed (from a high level) by nitrogen
starvation while the remainder is induced (from a low level).
Carbon
fixation
Morals of the Story
It’s possible in BioBIKE to
access information about
metabolic pathways
Morals of the Story
It’s possible in BioBIKE to
combine this information with
microarray data
Morals of the Story
No tool ever does exactly
what you want
Morals of the Story
No tool ever does exactly
what you want
In BioBIKE it’s possible to
make your own tools.
Users can package useful operations like the one shown
above into one-word functions and simplify their own
personal language. Functions that are particularly useful
may be incorporated into the general language.
Morals of the Story
And in fact, the language WAS
changed to make the process
simpler.
The same result may now be
obtained as shown below
Collaborators
Johnny Casey (Sequoia Cons.)
Sarah Cousins
Jeff Elhai (VCU)
Sahel Farhangi
Michiko Kato (now UC Davis)
JP Massar (Berkeley)
James Mastros (now Philip Morris)
Bogdan Mihai (VCU)
John Myers (Sequoia Cons.)
Nihar Sheth (VCU)
Jeff Shrager (CollabRx)
Mark Slupesky
Arnaud Taton (VCU)
Hien Truong (VCU)
Andy Whittam (Washington & Jefferson)
Development of BioBIKE was funded by grant DBI-0516378 from the National Science Foundation
Contact Jeff Elhai, Center for the Study of Biological Complexity, Virginia Commonwealth University
(E-mail) [email protected], (Tel) 804-828-0794