SEISMICgraph: Web App Documentation

Overview

SEISMICgraph is a webapp that enables researchers to utilize the seismic-graph Python package without learning the requisite coding skills.

SEISMICgraph accepts mutational profile data from SEISMIC-RNA, ShapeMapper2, or RNA Framework, and generates plots and analysis of that data.

Workflow for SEISMICgraph

The workflow for SEISMICgraph is as follows:

  1. Upload a dataset, or use the sample dataset.

  2. Make Data Selections:

    • Filter the data by making selections for which parts of the dataset to use.

    • As you filter the dataset, you will see that the number of Rows Selected updates with your adjustments. A row is an entry in the dataset, and it can be uniquely identified by its Sample, Reference, Section, and Cluster.

  3. Plot Selections:

    • Choose what type of plot to generate from the pre-existing plot types.

    • Some plots require a specific number of rows to be selected. For example, Mutation Fraction requires a single entry to be selected. To select a single row, choose values for Sample, Reference, Section, and Cluster to uniquely identify the row of interest.

    • Some plots have options that are specific to them.

  4. Plot:

    • The plot will generate. If the plot could not be generated, a message will be displayed indicating why it could not be generated.

Let’s go through these steps using the sample dataset as an example.

SEISMICgraph: Step by Step walkthrough

Upload Dataset

In this example, we will use the Sample Dataset. In the light blue panel, choose Use Sample Dataset. The sample dataset will be loaded into the app. This usually takes about 15 seconds.

Once the dataset has been loaded, we will see a few changes to the interface: under Data Selections we now see Rows Selected: 1401, and Section and Cluster have both been populated with values:

  • Section: 20-151

  • Cluster: average

Filter Dataset using Data Selections

For the filters Sample, Reference, Section, and Cluster, if there is only one value for that filter given the other filters, then that value will be automatically selected. For example, this dataset does not have clustering. The only Cluster is average, indicating the population average mutation rate. Since every selected row has the same value for Cluster, that value is displayed as the selection for Cluster, to make it clear that all selected rows are using population average.

For this example, we will generate a plot Mutation Fraction. Mutation Fraction displays the proportion of reads of a base that were mutations for every base in the reference sequence. It requires a single row to be selected, so we will use Sample and Reference to uniquely identify the row of interest.

By selecting the Reference of interest and clicking outside of the dropdown menu, the selection is sent, and when the loading indicator goes away, the Rows Selected updates to reflect the new selection. Once a Sample and Reference are selected, we can see that the dataset has been filtered down to one row.

Note: Datasets with multiple Sections and/or Clusters will require selections for those filters as well to uniquely identify a single row.

Make Plot Selections

  • Choose Mutation Fraction.

Plot

Click Plot. The plot will be generated in around 15 seconds. More computationally intensive plots like Number of Aligned Reads per Reference as a Frequency Distribution will take longer, especially on larger datasets.