GNF Malaria OPI Web Portal
Visit http://carrier.gnf.org/publications/Py, then follow the "P. y. & P. f. OPI Web Portal" link. This is an open-access web site.
Start your search from the overview page with gene or
cluster terms. Search results can direct you to either "Cluster Table
View" or "Gene Table View". Clicking on a cluster heatmap icon
opens the "Cluster Report". Each gene is linked to a "Cluster
Table View" containing all the clusters it is a member of. Each cluster is
linked to a "Gene Table View" containing all its gene members.
Enter gene locus name (e.g., PF11_0358), or cluster ID. Keywords in gene descriptions or cluster descriptions can also be searched. A cluster ID typically is assigned based on the source of the gene annotation; a cluster derived from a GO group uses its GO identifier (e.g., GO:0008173); a cluster derived from a publication uses its PubMed ID, e.g., GO:PM16497586; a cluster derived from genes curated or literature groups combined by GNF scientists has GNF as part of the identifier, e.g., GO:GNF0004.
Different types of search terms can be entered together and the "Smart Find" program has the intelligence of distinguishing what is a locus name, a cluster ID, or just a general keyword. Click the "Example" button, then the "Smart Find" button to check what the search engine can accommodate.
You can click the "Browse Gene Ontology" link to browse all the 156 malaria OPI clusters, click the "Browse Genes" link to browse all the P. f. and P. y. genes.
If you are interested in OPI clusters under a specific GO category, click the corresponding cluster count number (e.g., click "47" in the snapshot list all 47 clusters derived based on literature mining efforts).
Note: By
default, you are searching the malaria OPI result set (data set 1). In case you
are interested in browsing yeast or human OPI results, just choose another dataset
from the scrolling list. That enables you to search yeast/human genes and
clusters. Human genes are identified by their Affymetrix U133A probe set
identifiers.

Search results may contain both cluster entries and gene entries.
Each cluster entry is displayed with a clickable heatmap linked to its "Cluster Report" page, as well as a "Genes" link that will display all the genes in that cluster in the "Gene Table" view.
Each gene entry is displayed with a link to PlasmoDB, as well as a "GOs" link will display all the clusters that gene falls into in the "GO Table" view. The "Predictions" link shows all function assignments in a "Prediction Table View", where one can sort different parameters (sort by Rank by default).

Cluster report page contains all the details about the OPI cluster. It starts with key statistical information, followed by its gene list in a text area for the convenience of copying, then a link to the "Gene Table" view, its gene expression heatmap with sample description (mouse over), and a protein network graph if available.

Cluster
report page also contains a table listing all the gene members in the order of
their expression similarity to the representative profile, i.e., genes near the
top has higher correlation coefficients than genes in the bottom. Genes with a
are already annotated in the GO database,
a
means the prediction is not yet contained
in the GO database. The table also contains links to yeast/human orthologs, as
well as corresponding yeast/human OPI clusters in those datasets.
"Gene Table" view is an Excel-like web interface to browse a list of genes. If you reach this view by following the link provided by an OPI cluster, all genes are sorted by their Rank numbers by default. The web page contains many useful features as shown in the snapshot below.

"GO Table" view is an Excel-like web interface to browse a list of clusters. If you reach this view by following the link provided by a gene, all clusters are sorted by their LogP values, so that the cluster with the minimum LogP is displayed first. The web page contains many useful features as shown in the snapshot below.

"Prediction Table" view is an Excel-like web interface to browse a list of gene-cluster assignments. If you reach this view by following the "Predictions" link provided by a gene, all predictions are sorted by their Rank values by default. The web page contains many useful features as shown in the snapshot below. E.g., using a combined filter of "PY05002" in Gene_Name and "N" in inGO and Rank "<200", we can find the subset of predictions that assign new functions to PY05002. The confidence of prediction is determined by multiple factors, FDR, Correlation, Rank are just some of the parameters.

All data are available for download from our paper companion web site: http://carrier.gnf.org/publications/Py
Please email zhou at gnf dot org