<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<title>MOBY Use Case Panel Information</title>
<meta http-equiv="content-type"
content="text/html; charset=ISO-8859-1">
<meta name="author" content="Fiona Cunningham">
</head>
<body>
<h2>Scenario: </h2>
<br>
A researcher is doing a microarray experiment and want to both analyse his
data as well as compare it to similar experiments in the same or other species.<br>
<br>
<hr width="10%" align="left">
<h2> Background Knowledge: </h2>
<p>Non-affy microarray spot data, in its most simplistic form, could be represented
as three fields - Spot ID ("Reporter"), the label1 signal intensity, and
the label2 signal intensity. Generally speaking, a series of different
spots are observed in a single experiment and experiments are done over a
series (e.g. time course, or treatment regime). Thus a full microarray
experiment is, in its most simple case, a series of sets of spot data.
These data are generally clustered in some way, such that spots that
behave similarly, oppositely, or spots with functional similarity are grouped
together for visual inspection.<br>
</p>
<hr width="10%" align="left"><br>
<h1>Microarray Project Use Cases </h1>
<h2>Name: ClusterByFunction</h2>
<p><b>Scenario Reference: above</b></p>
<p><b>Problem: </b>Send a set of microarray data to be clustered by biological
function of the reporter spot<br>
</p>
<hr width="10%" align="left">
<h2> Primary Actors: </h2>
<ol>
<li>A biologist</li>
<li>A microaray data clustering service</li>
</ol>
<hr width="10%" align="left">
<h2>Other Actors: </h2>
a client program capable of visualizing microarray data<br>
<br>
<hr width="10%" align="left">
<h2>Initial State and Preconditions: </h2>
<p>The microarray data is unclustered. </p>
<hr width="10%" align="left">
<h2> End Result: </h2>
<p>The biologist recieves clustered data </p>
<hr width="10%" align="left">
<h2> Current Workflow: </h2>
<ol>
<li>data is imported into a local clustering tool, usually not web-based.</li>
<li>clustered data is visualized<br>
</li>
</ol>
<br>
<hr width="10%" align="left">
<h2> Existing Workflow Limitations: </h2>
<ol>
<li>not integrated with any other tool</li>
<li>available tool may not have all clustering options, or may have outdated
functional mapping</li>
<li>existing infrastructure (often the users desktop machine) might be
slow</li>
</ol>
<hr width="10%" align="left">
<h2>Existing Workflow Exemplars: </h2>
<ol>
<li>privacy - data is not transmitted over net</li>
<li>user has access to full documentation of the algorithm or program so
they know what is happening.</li>
</ol>
<br>
<hr width="10%" align="left">
<h2>MOBY Workflow: </h2>
<ol>
<li>available clustering services present themselves as soon as you have
microarray data in-hand</li>
<li>clustered data is returned</li>
</ol>
<ol>
</ol>
<hr width="10%" align="left">
<h2>MOBY Workflow Limitations: </h2>
<ol>
<li>data is passed over the net and might be proprietary or e.g. medical
and thus private</li>
<li>don't necessarily exactly know how the clustering is being done</li>
<li>generally large volumes of data are clustered, computationally intensive,
unlikely to find "free" services<br>
</li>
</ol>
<p><br>
</p>
<hr width="10%" align="left">
<h2> MOBY Workflow Exemplars: </h2>
<ol>
<li>service might have most up-to-date functional mapping</li>
<li>several services might present themselves who have greater computational
resources than the user himself<br>
</li>
</ol>
<p><br>
</p>
<hr width="10%" align="left">
<h2> Discussion: </h2>
<hr width="10%" align="left">
<h2> Priority:</h2>
<p>Desirable - if we can find anyone who is generous with their compute cycles
;-)<br>
<br>
</p>
<hr width="10%" align="left">
<h2> Key References: </h2>
<p><br>
<a href="%20http://www.genome.gov/page.cfm?pageID=10005336"></a></p>
<hr><br>
<br>
<h1>Microarray Project Use Cases </h1>
<h2>Name: FindSimilarExperiments</h2>
<p><b>Scenario Reference: above</b></p>
<p><b>Problem: </b>Discover if there are any experimental data publically
available from a similar experiment in another organism<br>
</p>
<hr width="10%" align="left">
<h2> Primary Actors: </h2>
<ol>
<li>A biologist</li>
<li>a public microarray data repository (e.g. SMD)</li>
</ol>
<hr width="10%" align="left">
<h2>Other Actors: </h2>
<br>
<br>
<hr width="10%" align="left">
<h2>Initial State and Preconditions: </h2>
<p>biologist has defined (somehow) the experimental conditions of interest
- I believe much of this type of info is stored in MAGE-ML, but I also think
there are other more explicit ontologies being built to describe experimental
conditions.<br>
<br>
</p>
<hr width="10%" align="left">
<h2> End Result: </h2>
<p>Biologist is informed of the availability of experimental data from studies
conducted under identical or similar experimental conditions </p>
<hr width="10%" align="left">
<h2> Current Workflow: </h2>
<ol>
<li>not possible? </li>
</ol>
<br>
<hr width="10%" align="left">
<h2> Existing Workflow Limitations: </h2>
<ol>
<li><br>
</li>
</ol>
<hr width="10%" align="left">
<h2>Existing Workflow Exemplars: </h2>
<ol>
<li><br>
</li>
</ol>
<br>
<hr width="10%" align="left">
<h2>MOBY Workflow: </h2>
<ol>
<li>an experiment-type matching service presents itself as`soon as you
have a MAGE-ML or other ontology descriptor</li>
<li>experiment ID's are presented to the user</li>
<li>microarray data associated with that experiment can be downloaded using
that ID</li>
<li><br>
</li>
</ol>
<hr width="10%" align="left">
<h2>MOBY Workflow Limitations: </h2>
<ol>
<li><br>
</li>
</ol>
<p><br>
</p>
<hr width="10%" align="left">
<h2> MOBY Workflow Exemplars: </h2>
<ol>
<li><br>
</li>
</ol>
<p><br>
</p>
<hr width="10%" align="left">
<h2> Discussion: </h2>
experimental type mapping will be tricky to get right... but useful!<br>
<br>
<hr width="10%" align="left">
<h2> Priority:</h2>
<p>Desirable<br>
<br>
</p>
<hr width="10%" align="left">
<h2> Key References: </h2>
<p><a href="http://genome-www5.stanford.edu/MicroArray/SMD/">Stanford Microarray
repository</a><br>
</p>
<hr>
<h1><br>
</h1>
<h1>Microarray Project Use Cases </h1>
<h2>Name: Retrieve Results of Homologous Loci</h2>
<p><b>Scenario Reference: above</b></p>
<p><b>Problem: </b>You have clustered your genes and have found a cluster
of interest where the members of the cluster appear to behave similarly to
the experimental conditions. Hypothesizing that these genes are co-regulated,
you wish to know if these genes in your organism, or the homologues in other
organisms, generally behave the same way under any other experimental conditions.<br>
</p>
<hr width="10%" align="left">
<h2> Primary Actors: </h2>
<ol>
<li>A biologist</li>
<li>a public microarray data repository (e.g. SMD)</li>
</ol>
<hr width="10%" align="left">
<h2>Other Actors: </h2>
<ol>
<li>a way to determine putative homologous genes - perhaps Blast, combined
with GO?<br>
</li>
</ol>
<br>
<hr width="10%" align="left">
<h2>Initial State and Preconditions: </h2>
<p>a set of loci names in-hand<br>
<br>
</p>
<hr width="10%" align="left">
<h2> End Result: </h2>
<p>step 1: a set of original & homologous loci, grouped by organism<br>
step 2: microarray data for any/selected experiments that have involved
those loci </p>
<hr width="10%" align="left">
<h2> Current Workflow: </h2>
<ol>
<li>know a priori or determine by sequence homology searching a list of
putative homologous genes from other organisms</li>
<li>go to http://genome-www5.stanford.edu/cgi-bin/SMD/cluster/QuerySetup.pl</li>
<li>select, one by one, various experiments and various organisms, or select
"all"</li>
<li>at the subsequent page http://genome-www5.stanford.edu/cgi-bin/SMD/cluster/QuerySetup.pl
select arrays one by one</li>
<li>at the subsequent page enter the gene list appropriate for that organism</li>
<li>download the microarray data<br>
</li>
</ol>
<br>
<hr width="10%" align="left">
<h2> Existing Workflow Limitations: </h2>
<ol>
<li>extensive manual clicking, cutting, and pasting<br>
</li>
</ol>
<hr width="10%" align="left">
<h2>Existing Workflow Exemplars: </h2>
<ol>
<li><br>
</li>
</ol>
<br>
<hr width="10%" align="left">
<h2>MOBY Workflow: </h2>
<ol>
<li>[numerous ways to retieve a list of homologues, either en masse, or
by organism]</li>
<li>submit gene list to microarray service</li>
<li>retrieve [all] microarray results for the submitted genes<br>
</li>
<li><br>
</li>
</ol>
<hr width="10%" align="left">
<h2>MOBY Workflow Limitations: </h2>
<ol>
<li>potential to get more data than you bargained for if we don't have
parameters!<br>
</li>
</ol>
<p><br>
</p>
<hr width="10%" align="left">
<h2> MOBY Workflow Exemplars: </h2>
<ol>
<li><br>
</li>
</ol>
<p><br>
</p>
<hr width="10%" align="left">
<h2> Discussion: </h2>
<br>
<br>
<hr width="10%" align="left">
<h2> Priority:</h2>
<p>Highly Desirable<br>
<br>
</p>
<hr width="10%" align="left">
<h2> Key References: </h2>
<p><a href="http://genome-www5.stanford.edu/MicroArray/SMD/">Stanford Microarray
repository</a><br>
</p>
<hr><br>
</body>
</html>