https://github.com/vjcitn/bioctechov
https://github.com/vjcitn/bioctechov
Last synced: 4 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/vjcitn/bioctechov
- Owner: vjcitn
- Created: 2018-12-24T12:58:44.000Z (over 6 years ago)
- Default Branch: master
- Last Pushed: 2018-12-24T14:17:04.000Z (over 6 years ago)
- Last Synced: 2025-01-09T13:46:19.831Z (5 months ago)
- Language: Makefile
- Size: 387 KB
- Stars: 0
- Watchers: 3
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
---
title: "A technical overview of Bioconductor, late 2018"
author: "Vincent J. Carey, stvjc at channing.harvard.edu"
date: "`r format(Sys.time(), '%B %d, %Y')`"
vignette: >
%\VignetteEngine{knitr::rmarkdown}
%\VignetteIndexEntry{}
%\VignetteEncoding{UTF-8}
output:
BiocStyle::html_document:
highlight: pygments
number_sections: yes
theme: united
toc: yes
---
# Introduction
Bioconductor is an open-source, open-development R-based software and data ecosystem for the analysis and comprehension of genome-scale assays. This document briefly describes technical concepts underlying the current system.
Three main lobes of Bioconductor are
- the IT infrastructure for source-code control, continuous integration, multiplatform deployment, release synchronization and archiving, construction of containers and VM images, markdown processing
- the domain-specific infrastructure that defines key user-facing classes and methods, such as Biostrings, SummarizedExperiment and GRanges, OrgDb/EnsDb SQLite-based annotation database schemas and interfaces, and the AnnotationHub/ExperimentHub resource discovery and caching methods
- the user-facing/developer-facing package ecosystem and support facilities
Current, lightly coordinated development efforts include
- lazy computation with out-of-memory array-like resources using DelayedArray
- efficient environment-adaptive computation of matrix decompositions
- general attack on massive single-cell assays
- API-driven resolution of queries on annotation (SRAdbV2 at github.com/seandavi) and matrix content (rhdf5client, restfulSE)
Larger infrastructure-oriented requirements will likely emerge through involvements with
- NHGRI AnVIL
- Human Cell Atlas
# IT infrastructure
Since its inception, Bioconductor has evolved in concert
with the R language, which itself has a bifurcated release
process.
- In R, the 'R-devel' branch incorporates new features
and code that may break aspects of the language.
The 'R-patched' branch evolves to fix bugs in the current
release of R, and no new features may be introduced.
- In Bioconductor, the same distinction is present,
but the branches are called bioc-devel, and bioc-release,
respectively. Furthermore, the bioc-devel branch transitions
to bioc-release every six months. When these transitions
occur, the version of R (patched vs. devel) that is used for developing
code also changes. Details on synchronization
are provided at the [developers advice page](https://bioconductor.org/developers/how-to/useDevel/).
## Build system for continuous integration
For a top-down approach to understanding the build system,
let's look at the landing page for the IRanges package.
All Bioconductor packages are accompanied by HTML pages
in this format.
- Badges describe the package portability (platform),
frequency of download (rank), public
discussion frequency (posts), longevity, current
build state (ok/warnings/error), and maintenance
activity (updated). These badges are updated daily.
- A DOI is computed for each package, referring to the
'release' version.
The landing page includes a collection of badges that report
high-level metadata about the package.
- The 'build' badge links to the build system reporting
page; on clicking 'warnings', we have
The actual warnings (which may be platform-dependent) are also
hyperlinked.
