nyu-molecular-pathology/ir-interpreter-flask

deprecated Flask version of IR-interpreter app

NOTE: This version of the app is deprecated, see the latest version here: https://github.com/NYU-Molecular-Pathology/IR-interpreter

• Demo site available here (not secured for clinical usage; for demonstration purpose only!)

IR-interpreter

Clinical interpreter for Ion Reporter and Oncomine analysis results

Overview

The Thermo Fisher S5 Next-Gen Sequencing platform can be used for the identification of structural variants, mutations, and other abnormalities in DNA and RNA samples using on-board Ion Reporter (IR) data analysis software suite. However, Ion Reporter does not include clinical interpretation of variants, which a desired feature for users of the platform.

The Precision Medicine Knowledge Base (PMKB) offers curated interpretations of many such variants which can be used for this purpose. Users can manually enter their desired variant into the PMKB website, manually look up entries in the PMKB Excel sheet available for download, or may use an API to query entries programmatically. While the latter option would be most ideal for a scripted solution, API usage requires the inclusion of Tumor and Tissue Type parameters, which are not always available for the program-user.

IR-interpreter provides an automated offline solution to this using the following method(s):

• creation of a portable SQLite database based on the PMKB Excel sheet (included)

• importation of user-exported IR .tsv formatted variant list

• matching of IR entries to PMKB entries with as many provided criteria as possible

• output of plain-text interpretations for each variant that are easy to copy/paste into systems such as PowerPath and EPIC

Usage

Installation

Clone this repo:

git clone --depth 1 https://github.com/NYU-Molecular-Pathology/IR-interpreter-flask
cd IR-interpreter

The included conda recipe can install all required dependencies in a fresh conda installation in the current directory:

make conda-install

• by default, all commands invoked through the included Makefile use this conda installation

• configurations for macOS and Linux are supported at this step

If you would prefer to manage dependency installation yourself, you can install required libraries from the included requirements.txt file.

Run Web App

You can start the web app with the Flask server with the following command:

make run

Make a Report

You can make a report from an Ion Reporter .tsv export file with the following command:

interpreter/report.py /path/to/Ion_Reporter_Sample.tsv

By default, output will be written to a file with the same path with the extension '.html'.

Filter by Tumor and Tissue type

You can provide a tumor or tissue type to filter interpretations by including extra arguments to the script:

interpreter/report.py /path/to/Ion_Reporter_Sample.tsv --tumorType "Adenocarcinoma" --tissueType "Lung"

An example can be run with the following command:

make demo

Monitored Directory

IR-interpreter can automatically process all input .tsv files in a directory with the included monitor.py module:

interpreter/monitor.py /path/to/inputDir

It can also be combined with rsync to first copy files from a directory on a remote server, process all inputs, then copy results back. A JSON formatted config file should be used for this (example .rsync.json included):

interpreter/monitor.py /path/to/inputDir --rsync --rsync-config /path/to/rsync.json

Filter by Tumor and Tissue type

You can enable interpretation filtering by tumor and tissue type by including a JSON formatted file with the same name adjacent to your input .tsv file. For example:

$ ls -1
Ion_Reporter_Sample.json
Ion_Reporter_Sample.tsv

$ cat Ion_Reporter_Sample.json
{
    "tumorType": "Adenocarcinoma",
    "tissueType": "Lung"
}

The JSON file (Ion_Reporter_Sample.json in this example) will be automatically detected by the monitor script and used to update filter criteria for interpretation output.

Automation

cron can be used to run this feature automatically. A sample crontab entry can be generated with

make crontab

And will look like this:

0 12,23 * * * . /ifs/home/kellys04/.bash_profile; cd /ifs/production/IR-interpreter; make monitor LOG=1 >/dev/null 2>&1

Timestamped logs will be deposited in the logs subdir in this directory.

Software

• Python 3.6+ (conda installation included for macOS and Linux)

Medical Disclaimer

This resource is intended for purely research purposes. It should not be used for emergencies or medical or professional advice. A finding of no interpretation does not necessarily indicate lack of relevance for any specific variant or biomarker alteration. Interpretations are not presented in ranked order of potential or predicted importance. These interpretations make no promise or guarantee of any clinical benefit (or lack of clinical benefit). Information contained on this website is not a substitute for a doctor's medical judgment or advice.