Ibex: Variational autoencoder for single-cell BCR sequencing.
Borcherding, N.; Sun, B.; DeNardo, D.; Brestoff, J.
Show abstract
SummaryB cells are critical for adaptive immunity and are governed by the recognition of an antigen by the B cell receptor (BCR), a process that drives a coordinated series of signaling events and modulation of various transcriptional programs. Single-cell RNA sequencing with paired BCR profiling could offer insights into numerous physiological and pathological processes. However, unlike the plethora of single-cell RNA analysis pipelines, computational tools that utilize single-cell BCR sequences for further analyses are not yet well developed. Here we report Ibex, which vectorizes the amino acid sequence of the complementarity-determining region 3 (cdr3) of the immunoglobulin heavy and light chains, allowing for unbiased dimensional reduction of B cells using their BCR repertoire. Ibex is implemented as an R package with integration into both the Seurat and Single-Cell Experiment framework, enabling the incorporation of this new analytic tool into many single-cell sequencing analytic workflows and multimodal experiments. Availability and ImplementationIbex is available as an R package at https://github.com/ncborcherding/Ibex. Reproducible code and data for the figure appearing in the manuscript are available at https://github.com/ncborcherding/Ibex.manuscript. A companion TCR-based approach is available at https://github.com/ncborcherding/Trex.
Matching journals
The top 2 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
Similar papers in this journal
- An unbiased comparison of immunoglobulin sequence aligners 93%
- Current challenges for epitope-agnostic TCR interaction prediction and a new perspective derived from image classification 92%
- Systematic evaluation of transcriptomics-based deconvolution methods and references using thousands of clinical samples 92%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.