Frequently Asked Questions

What is ORSO?

ORSO is the Online Resource for Social Omics. It is a network of next generation sequencing datasets that are connected on the basis of similarity between primary read coverage values and annotated metadata. These connections allow for discovery of datasets that may be important to your research interests. ORSO is a social network. Your interactions with data play an important part in shaping our network. Actions like favoriting a dataset or following a user create new links that influence how ORSO presents data to users.

What can ORSO do for me?

ORSO is a tool for discovering next generation sequencing data relevant to your interests. Each dataset hosted by ORSO is connected to the ORSO data network, where datasets are connected based on read coverage and metadata similarity. By selecting an individual dataset through the Experiment tab, you can find primary analytics about that dataset, such as relative gene coverage information, and a list of similar datasets. You can also explore the ORSO network in a top-down fashion through the Explore tab, enabling discovery of associated dataset groups.

Additional features become available once you create your own ORSO account. With an account, ORSO will make dataset recommendations tailored to you based on your history of favoriting experiments and following other users. You can also add datasets to the ORSO network. If set to 'Public', these datasets can be found by other users.

What data can I find on ORSO?

ORSO was launched hosting data from the ENCODE consortium, totaling more than 20,000 datasets. Users may also upload data for hosting by ORSO. The total number of datasets can be checked on the Overview page under the Explore tab.

In the future, we hope to add additional consortial data and to mine data repositories such as the GEO databank for public datasets. Watch the activity stream on your Home page for more details.

How can I add my own next generation sequencing dataset to ORSO?

For a dataset to be added, ORSO needs access to its read coverage information and metadata. Metadata is added by submitting a standard web-form. Regardless of the kind of next generation sequencing experiment, ORSO expects read coverage to be given in bigWig format. We recommend following the ENCODE consortium's protocols for generating bigWig signal files.

In order for ORSO to access coverage files, those files must be hosted and publicly available on a web server. This is similar to UCSC's requirements for bigWig hosting. If you do not have web hosting services available at your institute, UCSC provides guidelines about using third- party services to host your data. Once ready, you can provide the public URL in the experiment submission form.

What is the ORSO data network?

The ORSO data network is comprised of datasets, which make up the nodes, and similarities, which make up the edges. In short, it is a web of datasets where similar datasets are linked together. The ORSO web page is used to investigate and analyze the ORSO data network. Top-down views available under the Explore tab allow for the overall topology of the network to examined. Individual experiments may also be selected to generate lists of similar datasets.

How are datasets connected in the ORSO network?

Similar datasets in the ORSO network are linked together, so a connection in the ORSO network represents similarity between two datasets. ORSO uses two systems to create connections. The first considers metadata. We consider cell type and protein target when deciding whether or not two datasets are similar. We use the BRENDA Tissue and Enzyme Source Ontology and the STRING database to decide if two datasets are similar. If two cell types or protein targets share ontological parents or interaction partners, respectively, their associated datasets would be considered similar. The second system considers dataset read coverage values. To do this, we use a neural network classifier. This classifier is trained using coverage values from experiments with shared metadata information, such as the same cell type or protein target. After training, this classifier to applied to the read coverage values from two datasets to evaluate similarity.

User interactions also provide connections for the ORSO data network. By favoriting a dataset, similar experiments will be recommended to the user under the Experiment tab. By following a user, that user's personal datasets will also be recommended. Favoriting and following will also cause associated updates to be posted to your personal activity log on your Home page.

Where can I learn more?

Additional details can be found on ORSO's GitHub page. There, we include detailed documentation and the complete codebase for ORSO.