CSHL Computational Genomics - Reactome Workshopa


The Reactome Pathway database.


NAVIGATING REACTOME

  1. From the Sky to an event

    All of the events in Reactome are organized into a single diagram, the ReactionMap or sky, displayed near the top of most Reactome web pages. Each arrow in the sky corresponds to a single event. When the we browser is focused on a pathway, all of its component events are highlighted in the sky, and when it is focused on a physical entity, all events involving that entity as input, output, catalyst, or regulator are highlighted.

    Mouse over the sky, slowly. Note that names of events appear as the mouse moves over their arrows. Clicking on an event arrow takes you to the web page for that event.

  2. From the front page or TOC to an event

    The front page of the web site lists all the pathways annotated to date in Reactome. A expanded version of this list, showing sub-pathways, is on the page linked to the TOC button in the top menu bar.

    Clicking on an item in either list takes you to the pathway. Immediately under the sky panel on the pathway page, the whole pathway is shown in a hierarchical view. You can go directly to the page for any reaction shown there by clicking on its name.

  3. Navigating an event

    The Open Helix card shows a typical event page, highlighting its features, including links to other parts of Reactome and to external databases. Useful Reactome links include ones to


SEARCHING REACTOME

A simple search query can be made from the bar immediately under the top menu bar.

A much more powerful search is provided with the ExtendedSearch feature reached via the top menu bar. The lines in the form are connected by ANDs.

  1. For example, search for all events whose input is the exact phrase ATP, whose output is the exact phrase ADP, and whose compartment is the exact phrase cytosol. You should retrieve the names of 813 events in which something gets phosphorylated at the expense of cytosolic ATP. Restrict your search to events in the species Homo sapiens. You should get 106.

  2. Check the accuracy of the assertion that all curated reactions in Reactome are backed up by literature references. Look for reactions whose species is Homo sapiens, whose literatureReference attribute has no value, and whose inferredFrom attribute has no value. (That last restriction filters out human events manually inferred from experimental studies in other species.) Oops! A fair amount of our initial curation of DNA replication, the cell cycle, and intermediary metabolism was done from textbook sources, and while its likely to be accurate, it needs to be re-done and properly linked to experimental data.

ALTERNATIVE VIEWS OF WEB PAGES

By default, pages are displayed in a human-friendly EventBrowser format. Clicking on the Change default viewing format button at the bottom of the page pops open a menu that allows the user to switch to two legacy display formats (classic ) and also to switch to an InstanceBrowser view. This view is intended to display the underlying data structure of the page clearly.


PROJECTING DATA ONTO REACTOME

The SkyPainter feature allows users to overlay expression data onto the Reactome sky. It takes as input a list of protein sequence identifiers and, optionally, a numerical value corresponding to expression level, and generates an image in which each reaction arrow is colored according to the number of proteins involved in the event are on the input list, or the levels of expression of those proteins. Other analysis and display options are described on the SkyPainter page. Two sample data sets are provided. The identifiers with values data set is taken from a study of gene expression in HeLa cells proceeding synchronously through the cell division cycle, and the identifiers data set is a list of human proteins whose mutation is associated with disease states according to OMIM.


ORTHOLOGOUS PATHWAYS

This feature of Reactome projects curated human events on to model organisms whose genomic DNA sequences are known and hence for which full sets of gene models are available. Depending on the organism, the set of gene models might be extensively experimentally validated (brewers yeast Saccharomyces cerevisiae) or not (many prokaryotes). Putative orthologs are identified with the OrthoMCL strategy ([ref]). When model organism orthologs of all of the proteins involved in a human event (or 75% of the proteins involved in complexes) can be identified, a model organism ortholog of the event is inferred ([ref]).

Under each major pathway on the Reactome front page is a list of all organisms for which any events of that pathway have been inferred. Mousing over the lists of species names causes species-specific views of the sky to appear, showing only the events that have been inferred for that species. Clicking on the abbreviation for a species connects you to a page showing the inferred events in that pathway in that species. The biological meaning of these inferences is variable. The absence of events in Plasmodium falciparum corresponding to many parts of human nucleotide metabolism is a well known experimental result. The significance of the conservation of two events of apoptosis in all multicellular organisms examined, and Plasmodium, is less clear. Have structurally conserved proteins acquired divergent functions? Have ancient regulatory events been co-opted to modulate diverse processes? These inferred events are hypotheses, which one can test further through more detailed computational analyses of the proteins involved and through wet lab work.


DOWNLOADING MATERIAL

Software and bulk data can be downloaded click the download button on the top menu bar. Also, the contents of individual event and entity pages can be exported in SBML and BioPAX formats, useful for computational work and (soon) in PDF and RTF formats, useful for human readers, by clicking the hyperlinks at the bottoms of the event and entity pages.

ON-LINE USERS MANUAL AND DATA MODEL

These are accessible via the buttons on the top menu bar. Data model links to a narrative discussion of some key features of the Reactome data model; schema links to a detailed specification.


Course Home Page