IRIS

IRIS is a computational tool developed by the Xing lab for discovering alternative splicing (AS) derived immunotherapy targets. IRIS Explorer has two functions:

  1. The ‘Cell Line Selector’ tab aids in selecting cancer cell lines for experimental validation of antigens predicted by IRIS.

  2. The ‘Browse Data’ tab allows users to inspect events within TCGA and GTEx datasets through both splicing (percent spliced in) and junction count values.

Cell Lines

The web app allows users to inspect cell lines from three different sources: CCLE (n = 1019), Melanoma (n = 53), and PARCB (n = 15).

Required Inputs

The user is first required to select which cell line group to inspect (CCLE, Melanoma, or PARCB). Next, the user must input specific alternative splicing event(s) to inspect. The web app accepts two types of inputs: single AS event and file input.

For the single AS event input, the format for input is ac:gene:chr:strand:ES:EE:upEE:dnES

Note that this notation is also used in the IRIS output. Additionally, the user must specify the antigen producing junction of the event to probe: inc1, inc2, or skp (see figure below for details).

The web app also accepts a TSV file input to query for multiple AS events at one time which saves computation time. The first line of the TSV file must be “as_event antigen_junc” (where the two are tab separated). This first line is followed by AS events of interest. For each line, the file should have an AS event of the form ac:gene:chr:strand:ES:EE:upEE:dnES followed by the antigen producting splice junction for the event (either “inc1”, “inc2”, or “skp”). An example file can be downloaded from above the file input location. After the user uploads the file and presses “Query Event”, a dropdown selection widget will list all AS events in the file. The user must select an event of interest from the dropdown widget to further inspect.

Selecting Cell Lines

To aid in cell line selection, the web app provides various options to order, filter, and select attributes. The output table visualization displays all cell lines that match the user defined criteria. The following functions are supported by the web app:

Additionally, two interactive scatterplots are generated. The left scatterplot has normalized (counts per million) splice junction counts of the user-specified antigen producing junction for each cell line on the x-axis, and the right scatterplot has normalized (FPKM) gene expression values for the gene containing the AS event of interest on the x-axis. Both plots have PSI values of the AS event of interest on the y-axis.

Finally, there is a “Genome Browser” link under the scatterplots that allows the user to inspect the event range in the UCSC genome browser.

IRIS Browse Data

To aid in finding tumor-specific events, the “Browse Data” tab produces plots to help compare AS events between user-chosen sample groups. The user must provide a list of up to ten AS events to plot. After uploading a list of AS events, the user must choose sample groups (tumor and normal samples) to plot by checking the boxes next to the desired sample groups. Five plots will be generated for each AS event, and each plot will group the data according to the user-chosen groups:

The default definition for samples that express a junction is any sample that has a junction count of at least five for tumor samples and at least two for normal samples. The user can change this cutoff under “Minimum junction counts” before plotting.

Once the file of AS events has been uploaded and sample groups have been chosen, the user must press the ‘Plot’ button. After, a list of user provided AS events will appear in the form of a dropdown menu, and the user will be able to select a specific event to plot. After this event is selected, the five plots will be generated along with links to download the data required to make the plots.

Contacts

Beatrice Zhang beazhang@sas.upenn.edu
Yang Pan panyang@ucla.edu
Yi Xing Yi.Xing@pennmedicine.upenn.edu