February 24, 2022 -- The histone code concept needs to be redefined at the nucleosomal level, according to Michael-Christopher Keogh, PhD, chief scientific officer of EpiCypher. During a live webinar, Keogh shared information about the tools that are required to critically re-evaluate the genomic distribution and regulatory function(s) of histone post-translational modification (PTMs).
Below are the answers to questions from the live Q&A session during the webinar, provided by Keogh.
Is the SNAP-ChIP antibody testing data publicly available?
Data on 53xAbs to the H3K4 methyl states has been published in Molecular Cell. Extended data on >400 commercial antibodies to K-MetStat, K-AcylStat and K-OncoStat is in a manuscript that will be uploaded to bioRxiv / submitted for peer review in March 2022 (website already populated and will be announced at that time).
FOLLOW-UP: Did the SNAP-ChIP antibody data predict performance in CUT&RUN?
SNAP-ChIP was certainly a better predictor of CUTANA (CUT&RUN / CUT&TAG) performance than peptide arrays had proven for SNAP-ChIP. However, the discordance rate is higher than we're comfortable with. Our verdict? If you want to confirm performance in an approach perform in situ testing to standards -- no more surrogate assays.
Any recommendations for a pan-methyl antibody that can discriminate between me1, me2 and me3?
Not tested directly to that PTM type. We've used some Kub antibodies that purported to be pan- and they certainly aren't on Nucs (in preparation). Recent studies to characterize pan-target Abs tried degenerate peptide arrays and invariably find flanking sequence blind-spots. The presumption is that a polyclonal antibody (pAb) or pool of defined monoclonals (mAbs) that compensate for each other might have some use in some approaches. However, if the intent is to use for genomic mapping -- as above peptide arrays have proven to be unhelpful for predicting such ability so I'd not hold out much hope (and you'd certainly experience difficulty making an in situ standards panel).
You used salmon sperm DNA as competitors. Would using unmodified nucleosomes as competitors change your result?
For PWWP testing (where the domain can bind DNA and Kme/thus DNA-binding is the problem) you would likely get the same result with non-biotinylated competitor Nucs (though it would be a more expensive experiment). However, such competitor Nucs would likely work very well if you were exploring other entities: e.g. those that engaged the acidic-patch and a PTM.
You generated a lot of the data with CUT&RUN. Can you explain the difference between CUT&RUN and CUT&Tag?
We work with both immunotethering technologies and currently see CUT&RUN as the greatest all-rounder (can be used for every target: e.g. PTMs, readers, TFs and ATPases). CUT&Tag requires a high-salt wash step in the protocol and this could displace some Chromatin Associated Proteins (CAPs) - something we plan to test extensively as we launch our CUT&Tag kit in 2022. We believe CUT&Tag likely has the greatest future for specialized low-input applications.
You've shown you can do single to triple readers with dCypher. How far can you go? Can you do complexes?
The standard dCypher approach (on ALPHA beads) has been used extensively with up to two 'independent points of contact' (QUERY reader: TARGET nucleosomes) but then seems to saturate, so a modified approach was used to explore the trivalent PHIP. We've not yet tried to push it further and need some candidates (also we have another technique in R&D that will be explored in tandem). And yes -- the domains under study do not need to be in the same protein but can be complexed.
Have you done reader CUT&RUN with constructs other than BPTF? Is this approach broadly capable?
Now successfully done more than a dozen readers with great success. Thus far single domains can be pretty mediocre (not too surprised by that) unless you permit dimerization (e.g. via GST) or enforce it (by putting them in tandem). But paired domains (e.g. as in BPTF) can behave magnificently (this work to be submitted in coming days).
During the webinar, Keogh mentioned a preprint publication that the company was in the process of submitting, based on the topics presented. That article has since been published on bioRxiv on February 22.
Webinar Slides, courtesy of EpiCypher.