SBP_2019

SBP_2019: Sequencing Budget Optimization for scRNA-seq

SBP_2019 quantifies the tradeoff between sequencing depth (reads per cell) and number of cells in single-cell RNA sequencing (scRNA-seq) experiments to optimize estimation of multivariate generative models of gene expression.

Key Features:

• Tradeoff Quantification: Models the impact of sequencing depth and cell number on accuracy of multivariate gene expression inference.
• Variational Autoencoder Framework: Applies a variational autoencoder to perform error analysis of gene expression model estimation.
• Depth Threshold Analysis: Identifies diminishing returns beyond ~15,000 reads per cell and prioritizes increased read depth at shallow sequencing levels.
• FASTQ Subsampling and Processing: Implements a four-step workflow beginning with FASTQ subsampling and processing using kallisto bus.

Scientific Applications:

• scRNA-seq Experimental Design: Optimizes allocation of sequencing resources to improve clustering and multivariate analysis of gene expression in developmental biology, cancer research, and immunology.

Methodology:

SBP_2019 performs FASTQ subsampling followed by processing with kallisto bus, estimates multivariate generative models of gene expression, and applies a variational autoencoder-based error analysis to evaluate the effects of sequencing depth and cell number on model accuracy.