base2bio/compression repository overview

= Optimal compression scheme to balance ratio and decompression speed Jeremy Volkening :source-highlighter: pygments

[abstract] .Abstract Some resources spent most of their time in storage but also require quick access when needed. For these resources, compression speed is of little importance, as they are write-once-read-often objects. We tested a number of popular compression tools to look for one which provided fast decompression with reasonable file sizes.

== Methods

A variety of paired-end Illumina FASTQ datasets were used as input (TODO: finish description).

=== Software versions

[%header,cols="2*^",format=tsv] |=== include::versions.tsv[] |===

== Results

=== Compression ratio

.Compression ratio (higher is better) image::figure/fig_ratios-1.svg[Compression ratio (higher is better),align=center]

.Ratio of compressed files with Q8 binning (unbinned:binned, higher is better) image::figure/fig_ratios_q8-1.svg[Ratio of compressed files with Q8 binning (unbinned:binned, higher is better),align=center]

=== (De)compression speed

.Compression rate (MB/s) (higher is better) image::figure/fig_ctime-1.svg[Compression rate (MB/s) (higher is better),align=center]

.Decompression rate (MB/s) (higher is better) image::figure/fig_dtime-1.svg[Decompression rate (MB/s) (higher is better),align=center]

=== Memory usage

.Compression maximum resident memory image::figure/fig_cmem-1.svg[Compression maximum resident memory,align=center]

.Decompression maximum resident memory image::figure/fig_dmem-1.svg[Decompression maximum resident memory,align=center]

=== CPU usage

.Compression average CPU utilization image::figure/fig_ccpu-1.svg[Compression average CPU utilization,align=center]

.Decompression average CPU utilization image::figure/fig_dcpu-1.svg[Decompression average CPU utilization,align=center]