Hi jinyu ,
In a single sentence, the answer to your question can be summarized as, "the resolution of a genomic technology refers to the smallest genomic variant that can be assayed by the technology". For Single-base, then, it refers to a technology that can distinguish the presence and absence of SNVs reliably.
Part of the reason there are so many genomic technologies is the desire to cheaply assay classes of variation that affect nucleotide sequences over 9 orders of magnitude in size. Consider - a whole genome duplication event is of interest as an early step in carcinogenesis - thats on the order of 10^9. NGS tech is virtually blind to pure WGD events - but a much much older technology, karyotyping, can assay it easily. At 10^0, we have the SNV, which you allude to in the original post. Every OoM between those has scientifically and medically relevant questions that depend on it, too.
Nowadays since people are used to conceptualizing technologies that sequence every base the term can require clarification - its true. But we just need to remember that not all technologies work by giving information on individual DNA (or RNA) bases. However, this was untrue for ~1/2 the history of genomics.
- Consider: Does Karyotyping give you sequence level info? Nope. Is it a genomic technology? Yep. (Resolution: ~1 x 10^6 bases)
- FISH makes probes that bind DNA for the purpose of telling you whether a rearrangement has occurred (Resolution: ~1 x 10^4 bases)
- Optical Genome Mapping works by binding small probes to lil sequences here and there that overall give you a map of the large-scale structure of the genome so you can see if there are large scale changes (Resolution: ~5 x 10^2 bases)
- Genomax has summarized the technologies you mentioned in his comment, above - I will defer to his expertise there.
- Finally, there are technologies aimed at assaying single nt changes (Resolution: ~1 x 10^0 bp). These include, for instance, sequencing technologies, which work by tying a readout of some kind to (every) consecutive base in a linear string - such that a readout of primary structure is obtained from the technology - are said to provide "single base resolution". There are other non-sequencing techniques that do the same thing - DNA microarrays have probes designed to single nucleotide changes, such that the genotype of huge numbers of SNVs can be assayed in parallel. These don't generate sequence info, but do provide single base res. at certain positions in the genome.
An intriguing comment that you've made has to do with a perceived division between utility of single base resolution in DNA vs. in RNA. If I understand correctly, I think I'd frame it differently - that it has less to do with nucleotide vs. nucleoside and everything to do with the scale of analysis that is of interest to the experimenter.
The phenomenon of RNA-editing has led to issues with accurate sequencing of RNA, for example, in routine RNA-seq studies. For better or for worse, much of the RNA data generated was used to do things like differential expression analysis. By contrast, the number of attempts to report and/or directly analyzing primary sequence of RNA is comparatively small (to clarify, I am saying the data are analyzed at the level of belonging to a gene or transcript, rather than analyzed at an individual base level). But here again the issue is the desired resolution of the analysis, not what is possible/impossible with RNA. Its certainly possible and useful to do it, it just requires more care to get it right.
Sequencing by hybridization (https://en.wikipedia.org/wiki/Sequencing_by_hybridization ) is likely not giving you single base resolution. But sequencing by synthesis (most current sequencing technologies with exception of Nanopore) should offer single base resolution.
GROseq and eCLIP-seq are techniques that are used for mapping binding sites for RNA pol-II and RBP. But the final sequencing of the cDNA is done using seq by synthesis.