RASE - Resistance-Associated Sequence Elements

This repository contains the core RASE package. Other components of the software include the RASE pipeline and the RASE DB skeleton. For more information, see the RASE paper and the RASE supplementary materials.

Content

• Introduction
• Installation
• File formats
  • Prediction output (timeline)
  • Prediction output (snapshot)
  • RASE DB metadata
  • RASE/BAM
• Related repositories
• License
• Contact

Introduction

This repository contains the core RASE software for rapid inference of antibiotic resistance and susceptibility using genomic neighbor typing.

Installation

1. Setup an environment

2. Install RASE using PIP

   pip install --upgrade git+https://github.com/c2-d2/rase

File formats

Prediction output (timeline)

Tab-separated text file with the following columns:

column	description
datetime	datetime of the read
reads	number of processed reads
kbps	number of processed bps (thousands)
kkmers	number of matched k-mers (thousands)
ks	k-mer score (proportion of matched k-mers)
ls	lineage score
ls_pass	lineage score interpretation, 1=passing 0=failing
ln, alt_ln	predicted and alternative lineage
bm, bm_{feature}	best-matching strain (nearest neighbor) and its features
{ant}_ss	susceptibility score for the antibiotic {ant}
{ant}_pred	prediction (score interpretation): S=susceptible, R=non-susceptible, S! and R!=low confidence calls
{ant}_bm	resistance information for the best match, format: cat (mic)

See an example file.

Prediction output (snapshot)

Tab-separated text file with the following columns:

column	description
taxid	taxid of a database strain, _unassigned_ for reads without any k-mer matches with the database
lineage	lineage
weight	weight (cumulative "number of k-mers matches divided by the the number of best matches")
weight_norm	normalized weight
length	cumulative "read length divided by number of matches"
count	cumulative "count divided by number of matches"

See an example file.

RASE DB metadata

Tab-separated text file with the following columns:

column	description
taxid	taxid of a database strain
lineage	lineage
order	order for plotting
{feature}	arbitrary features (e.g., serotype or MLST)
{ant}_mic	original MIC string (e.g., <0.03)
{ant}_int	extracted MIC interval (0-0.03)
{ant}_cat	resistance category (S/R/s/r)

See an example file. Metadata files are generated in dedicated repositories (see RASE DB skeleton, S. pneumoniae RASE DB, and N. gonorrhoeae RASE DB).

RASE/BAM

File following the SAM/BAM specification, containing the following information:

field	description
QNAME	read name (ideally with the timestamp encoded)
FLAG	4 for unassigned, 0 otherwise
RNAME	strain taxid or name of an internal node
ln:i:	read length
h1:i:	number of k-mer matches at this node

Related repositories

• RASE supplementary. Supplementary Materials for the RASE paper, including figures, tables, experiments, and links to other related repositories.
• RASE DB skeleton. Skeleton for creating novel RASE databases.

License

MIT.

Contact

Karel Brinda <[email protected]>