With all the “LOL”s, “IMHO”s, and “YOLO”s out there, it’s like we’re all speaking in code. When we get stumped on what these things mean, we can always turn to Google. But what if you want to decipher the hypothetical “epigenetic code?” Well, one researcher has the answer—EMdeCODE, a new algorithm that does just that for histones. And he found a new role for H3R2me1 in the process.
EMdeCODE uses peaks from ChIPSeq data using a separate peak caller as the input so that it can handle data from different peak calling algorithms and from many experiments. And that way, if a better ChIPSeq analysis tool comes along, it’s not obsolete.
The algorithm takes histone mark profiles and converts them into a new “supermark” tied to a function. It outperformed similar tools and had better precision and sensitivity. Here’s a little more info:
- EMdeCODE successfully identified enhancer elements. The “enhancer code” featured lots of H3K36ac (known to be at enhancers), H2bK120, H3K4ac, among other modifications.
- The algorithm also predicted retroviral integration sites, which include marks previously recognized as active transcription and open chromatin modifications.
- With EMdeCODE, he could tell coding from noncoding regions. People had thought that epigenetics couldn’t distinguish the two, but turns out, that was only true for the area near the promoter. Typical transcription marks like H2aZ, H3K4me3, and H3K36me1-3 were enriched within coding regions.
Surprisingly, H3R2me1 was in coding gene bodies, but absent from noncoding regions (like for lncRNA). Until now, no one has really known what H3R2me1 does, says the author, though others have linked it to euchromatin maintenance.
He says that all of these data show that H3R2 maintains and protects coding regions.
See for yourself at Nucleic Acids Research, December 2012.