Design notes

[jayhesselberth]

Data structures

• BAM, sequencing_summary per run.
• Calculate BCerror from BAM, FASTA.
• Use bioconductor data structure to keep everything together, SummarizedExperiment or RangedSummarizedExperiment (?).

Methods

• QC. per-sample metrics
• BCerror. Straightforward with cpp11? Otherwise https://github.com/rnabioco/detectrms-rs
• Comparison. Two samples, differential expression & modification.
• ability to interact with pod5

Plotting.

• tRNA heatmap.
• tRNA roadkill. maybe like https://github.com/JPSieg/R2easyR

[lkwhite]

Would be nice to eventually include some functionality for looking at intron containing pre-tRNAs and their modifications. This would enable insight into modifications deposited before / after intron removal. Also would encourage alignment to a reference containing these molecules — I haven't explicitly tested how our alignment and filtering conditions perform on intron containing tRNAs but my guess would be they make it through alignment and contribute to bcerror background noise as indels in the anticodon loop.

[lkwhite]

Another eventual nice to have would be to be able to optionally annotate the canonical long range base-pairing contacts formed in the 3D structure in 2D roadkill plots.