~todo~

[Alexander-Morin]

General setup

• Common genomic tables (protein coding, blacklist...) across projects need package
• Util functions needs a review
• Util functions -> package with TR agg functions?
• Similarly, binding specificity call is largely copy pasted from TR agg
• Formalize DHS download
• Saving GRanges and getting npeaks for meta could potentially be bundled
• More explicit about de-dup meta
• Review slide of Unibind/DAMO pipeline

GRanges analysis

• Reduce duplicate experiments
• Frequent bound genes have dispersed peaks?
• Motif analysis: motifs in DHSs lacking binding (top candidates not represented in Unibind?)
• Motif analysis: candidate co-factors in freq bound regions specific for a TF
• Look at UK Biobank depletion ranks (focus on TF-specific regions?)
• Look at ultra-conserved regions
• Notable regions lacking cCRE status?

Binding score analysis

• Formalize frequently/never bound
• Compare mouse dev transcriptome and HPA with infreq/freq bound genes
• Formalize clustering of TFs
• Binding specificity model on TF families
• Potentially conserved TF-target interactions
• bind summary needs to be more explicit about choice of permissive/robust and raw mat (for model) vs QNL (for mean)

Misc

• Compare robust and permissive needs to be updated
• ENCODE pipeline output vs Unibind for batch 1 TFs needs to be updated