Methylation | WGBS | • No a priori sequence selection | • High cost and may require higher coverage • cannot distinguish type of modification at cytosines | Y |
Methylation | RRBS | • Lower cost | • Limited mainly to CpG islands • cannot distinguish type of modification at cytosines | Y |
Protein Localization | ChIP-seq | • Genome-wide profiling of histone modifications and DNA-protein interactions (Histone H3 acetylation/methylation, TF binding site identification, SE identification) | • Survey only one type of interaction (protein) at once • lots of sources of noise/bias • requires good antibodies • requires input DNA and isotype controls • requires large input of cells | Y |
Protein Localization | CUT&RUN | • Fewer input cells required than ChIP • less noise • fewer sequencing reads required • no cross-linking required | • Requires good antibody • potential for overdigesting DNA | Y (CUT&Tag, uliCUT&RUN) |
Chromatin Accessibility | DNAse-seq | • Identify a range of cis and trans regulatory elements including TF binding sites | • High input cells requirement • more time-consuming that ATAC • sequence bias | Y |
Chromatin Accessibility | ATAC-seq | • Identify a range of cis and trans regulatory elements including TF binding sites • minimal input cells required • increased sensitivity over DNAse-seq • simple protocol | • Footprint profiles can be less well-defined than DNAse-seq • potential mitochondrial DNA contamination | Y |
Chromatin Accessibility | MNAse-seq | • Nucleosome occupancy and positioning • can be used to predict higher-order structure (e.g. 3D) | • Requires crosslinking • highly dependent on enzyme concentration • some sequence bias | Y |
Chromatin Accessibility | FAIRE-seq | • No sequence bias • simple protocol • no enzymes required | • Requires crosslinking • lower resolution (crosslinking binds chromatin but also TFs) • large input cell requirement | N |
3D Conformation | 3C (Chromosome Conformation Capture) | • Identify single chromosomal interaction (one vs. one) | • limited resolution (by 6bp cutters) • laborious • PCR biases • high library complexity • single viewpoint | N |
3D Conformation | 4C (Circular 3C) | • Improved resolution over 3C • can identify very long range interactions • can identify all contacts for a locus (one vs. all) | • Biases from circularization • PCR biases • high input cell requirements • single viewpoint | N |
3D Conformation | 5C (3C Carbon Copy) | • Can identify many contacts for multiple loci (many vs. many) | • Bias introduced by probe ligation efficiencies • not all fragments can bind probes • all vs. all prohibitively expensive | N |
3D Conformation | NG Capture-C | • Analyze hundreds of viewpoints • can identify PCR duplicates (low bias) • highest sensitivity and resolution • fewer input cells required | • Occasional non-specific interactions | N |
3D Conformation | Hi-C | • Maps contacts across whole genome (all vs. all) • kilobase resolution | • Fewer contacts per fragment than 4C or Capture-C • higher resolution versions require extremely high sequencing depths | Y |
3D Conformation and Protein Localization | ChIA-PET | • Combines 3D interactions with protein interactions | • Interactions defined by few reads • high input requirements • bias toward interactions with targeted protein | Y* (ChIA-Drop: single molecule, Zheng et al., 2019) |
3D Conformation and Protein Localization | Hi-ChIP (and PLAC-seq) | • Lower input required • higher yield than ChIA-PET • higher signal to noise over Hi-C | • Bias toward interactions with targeted protein | N |