by Janie
Recently I stumbled across Simon Berger's shattered glass artwork, and was reminded of RNA-seq and transcriptomics, of all things.
The Swiss artist's pieces each begin with a large pane of glass. He shatters the pane, sometimes by heavy blows via hammer that generate large fragments, sometimes with more precise taps that produce concentrated clusters of lacerations, and yet other times by scoring the glass with a blade and then snapping it along that axis. Varying the method of glass-shattering serves as this artist's palette, from broad and linear fractures to fine spiderweb-like veining. Some of his finished works are single panes of glass, like his portrait of Kamala Harris, while others are multiple panes stacked on top of each other, like his "Shattering Beauty" (Fig. 1).
Figure 1. "Shattering Beauty" by Simon Berger. Source
Simply imagining raising a hammer to again strike the glass pane of a work in progress, I feel anxious – such a fine line between continuing to make art and needing to start over! What Berger does seems like such a delicate balance between creation and destruction, where many little lines make up the whole, so that the final "picture" we see is our subjective interpretation of the amalgamation of those many data points. Sound like a transcriptome yet? (There is also the fact that Berger refers to his glass-shattering art-making process as "morphogenesis," a term that of course also refers broadly to biological phenomena that have been the subject of transcriptomic scrutiny across various organisms.)
Fig. 2 In vivo RNA 3' end profiles reveal that the majority of transcripts are not full-length in E. coli). Φend indicates the fraction of 3’ end sequencing reads that map to the 3’ end of the mature transcript (the lower this value, the more transcript fragments there are). Left: the Φend calculated for various transcripts, as correlated between two replicates. Right: the correlation between the fraction of full-length transcripts as calculated by 3’ end-seq and by Northern blot for a select few transcripts. From Figure 1 of Source.
It was Berger's piece "Shattering Beauty" that reminded me of a paper on transcript fragments from a couple years ago. In 2022, Herzel et al. reported that a substantial portion of the E. coli transcriptome consists of partial transcripts arising from mRNA decay. Around 60% of the transcripts studied were truncated at the 3’ end. These partial transcripts were not primarily nascent RNAs, as indicated by the very different 3’-end-sequencing profiles of total RNA and of nascent RNA. The notion of the fragments arising from decay was also supported by correlations between these internal 3’ end sites and two degradation signatures – cleavage sites for RNase E (the main endonuclease in mRNA decay) and polyadenylation – in both wild type and a PNPase mutant (the main exonuclease in mRNA decay, which comes into play after RNase E). What was most interesting, when thinking about interpreting data as real biology, was that the majority of translation was occurring on full-length and nascent transcripts, not on these degraded and degrading fragments. In other words, the pretty picture of a cell's inner workings painted by RNA-seq data might not necessarily be a very biologically relevant portrait.
This is not to say that all RNA degradation is not biologically relevant. Some cases of mRNA decay are "strategic." The E. coli malEFG operon comes to mind. The 5' end of its transcript, which harbors the gene encoding the maltose-binding protein, is protected against exonuclease activity by REP sequence hairpins, while the unprotected 3' end of the transcript is more easily degraded. This differential degradation along the same transcript results in higher levels of that maltose-binding protein, which is, fittingly, required in higher levels than the other proteins encoded by the rest of the operon. Similar principles apply to polycistronic transcripts of other ABC transport systems, on which hairpin structures confer resistance against PNPase activity for upstream genes.
It's also worth noting that the origin of transcript fragments also depends on the species. While the partial transcripts in Herzel et al.'s E. coli were primarily a result of mRNA decay, a recent publication by Ju et al.demonstrated that this is not the case in Mycobacterium tuberculosis. In Mycobacterium, the majority of partial transcripts are in fact nascent mRNAs, as suggested by their frequent association with RNA polymerases.
Why exactly this paper along with Herzel et al. jumped to mind upon seeing Berger's "Shattering Beauty" is probably some combination of 1. Fragments of both RNA and glass; 2. A shared sense of "you see what you want to see": an assemblage of glass fragmentation patterns can look like a skull if you choose to look at it from one orientation. This itself smells of RNA-seq data interpretation and the pitfalls thereof (hence the skull); 3. An unconscious linkage of glass -> transparent -> water -> cell, and of skull -> death -> decay -> RNA degradation; 4. The image of shattering glass is reminiscent of busting open cells to extract their RNA. Transcriptomics seeks answers to life via the death of cells (hence the skull, again); 5. The side view of "Shattering Beauty" looks like a pile-up of mapped reads.
I do like the idea of pairing artworks with scientific bodies of work. I'd love to know what other pairings others have come across! In the meantime, I'll be keeping an eye out myself for more.
Do you want to comment on this post? We would be happy about it! Please comment on Mastodon, Bluesky, or on 𝕏 (formerly Twitter).
Comments