Declarative Bioinformatics (2019-08-05) (Work in Progress)

This post will attempt to describe and explore a “declarative programming interface” for bioinformatics. The specific purpose of this declarative interface is to determine the necessary subset of relational algebra, relational calculus, or other language that a storage system would need to provide to support various bioinformatics use cases. I will start with gene expression and single cell RNA sequencing (scRNA-seq) use cases, but I will explore how this interface generalizes to other bioinformatics use cases.

Bioinformatics is a broad field that tries to understand molecular and cellular biology through the analysis of biological data. Taking a top-down approach to determining a declarative programming interface, I start with some key biological entities and concepts that exist in gene expression analysis and single cell RNA sequencing.

List of Biological Entities and Concepts (Work in Progress)

Molecular and cellular biology terminology:

• subpopulation
• biological processes
• gene expression
• gene:
  • comes from RNA
  • a gene corresponds to a particular genomic region that codes for a protein
  • isoforms based on alternate splicing
• marker gene
• orthologous genes
• gene set - …
• peripheral blood mononuclear cell (PBMC)
• protein:
  • unit of work in biological mechanisms
  • participates in various pathways
• pathway:
  • lineage of protein activity preceding a particular biological mechanism

Sequencing terminology:

• sequencing run
• sequencing library
• sequencing read paired-end
• primer
• barcoded, full-length cDNA
• signal-to-noise ratio of scRNA-Seq:
  • signal - optical data from the reaction of interest
  • noise - obfuscation of signal from:
    • accumulated or lingering reagents
    • …

Analysis terminology:

• enrichment analysis-based approach
• cellular state - a representation of cell life-cycle and function based on protein activity
• batch effect - …
• drop-out event - …
• pipeline
• assay
• high-throughput assay
• transcript-level data
• within-tissue and cross-tissue

Ontology terminology:

• GO-BP

Resources

• Cell Trajectories - A comparison of single-cell trajectory inference methods
• Construction of Single Cell Sequencing Libraries - 10x Chromium GEM Library Construction
  • Pages 14 - 16; Section “Stepwise Objectives”
• Single Cell Sequencing Analysis - Massively parallel digital transcriptional profiling of single cells
  • Page 3; Figure 1
  • Page 5; Figure 3
  • Pages 4 - 6; Section Subpopulation discovery from a large immune population