Why SEISMIC-RNA

We built SEISMIC-RNA because none of the existing software for mutational profiling data met all our needs. A piece of software compatible with Linux or macOS that could analyze multiple samples (FASTQ files) from many transcripts simultaneously, in parallel. One that would perform all steps of data processing, from alignment to structure prediction, autonomously with a single command. And that would correct known biases in DMS-MaPseq and SHAPE-MaP data, applying rigorous filters to ensure only high-quality input data wound up in the results.

From these requirements, we designed SEISMIC-RNA to be both easy to use yet highly customizable, and both speedy yet accurate. You can install it via pip and conda, then run it with a single command (seismic wf) that performs all steps from alignment to structure prediction using default settings that work in most cases. For more advanced usage, you can run steps individually and customize dozens of options to best suit your data. And if the command-line interface doesn’t do everything you need, SEISMIC-RNA also provides a Python API (import seismicrna) for access to all its classes and functions. We hope that by having a low floor and high ceiling, SEISMIC-RNA will make it easy to investigate the structure ensembles into which your favorite RNAs fold.

What SEISMIC-RNA does

Given sequencing reads from a chemical probing experiment (such as DMS-MaPseq or SHAPE-MaP) and one or more reference sequences, SEISMIC-RNA:

  1. Optionally demultiplexes the FASTQ files.

  2. Aligns the reads to the references.

  3. Identifies the mutations in each read.

  4. Filters reads and positions on user-specified criteria.

  5. Clusters reads by their mutations to infer structure ensembles.

  6. Predicts secondary structure(s) using the mutation rates.

  7. Generates a variety of graphs and RNA structure diagrams.

The same pipeline runs from a single wf command or step-by-step from any intermediate point.

What distinguishes SEISMIC-RNA

End-to-end and reproducible

The entire workflow above can be run with one command, seismic wf, removing the need to write a script with multiple commands. Every step saves its output files so that the workflow can be resumed from the middle if needed.

Scales to many samples and transcripts

SEISMIC-RNA searches directories recursively for input files, allowing you to provide many input files at once. It then processes input files in parallel to maximize speed.

Automated organization

SEISMIC-RNA automatically organizes files using the names of the given samples and transcripts, sidestepping the need for manual organization.

Structure ensemble inference

SEISMIC-RNA’s clustering step determines whether the RNA folds into multiple structures, which is essential for modeling RNA structures accurately.

Scales to long RNAs

SEISMIC-RNA can handle both short amplicons and transcripts tens of thousands of nucleotides long.

Open-source and scriptable

SEISMIC-RNA is released under the GPL (see License) and exposes both a CLI (see Command Line Reference) and a Python API (see API Reference).

Where to go next