Skip to main content

Biological News

You Don’t Understand Long Noncoding RNAs

Let’s confuse ourselves thoroughly and talk about long noncoding RNA. It’s just part of the menagerie of RNA species that have come to light over the last twenty years, of course, and honestly you can confuse yourself with any of them. But just asking “what are lncRNAs doing?” is enough right there.

They’re lengthy stretches (>200 bases, as an arbitrary cutoff) of RNA that don’t seem to be associated with any particular gene, but are nonetheless transcribed and turned loose to float around. If you look closely, you can find a great number of these things (on the order of tens of thousands) inside cells. The assumption is that they have to be doing something, but what? They tend to be rather cell-type specific, and even cell-cycle or developmental-stage specific, which definitely argues for their utility. They don’t seem to be translated into protein, but even that is the subject of argument, with some groups saying that there are so-called noncoding RNAs that have been misannotated, and other proposals that they’re somehow translated into short-lived peptides that are difficult to detect. But it’s for sure that many of these things don’t look like they have open reading frames in them.

Are they conserved? More arguing. My understanding is that at a sequence level, they do seem to be more so than just a bunch of random RNA, for sure – people have seen what looks like conservation across species, for example. But it’s been hard to show that these variants are doing the same thing (whatever that may be) because the levels of the different transcripts vary so much (you’d think that a conserved type would show up in roughly the same amounts, but that doesn’t seem to be the case).

And as for relevance to disease, prepare yourself if you do a search: there’s an absolute firehose of literature on possible connections to cardiovascular disease, various tumor types, CNS conditions, developmental pathways of all kinds, and pretty much everything else you can think of. That’s what having no well-defined mechanism will do for you – lncRNAs are a blank screen on which you can draw most anything. Effects on mRNAs, on oligonucleotide stability or localization, on chromatin or the assembly of transcriptional machinery – sure, whatever. And the number of papers is growing steadily with no sign of leveling off. I would very much want to avoid having to bring order to even a small part of the literature in this area – there are so many reports, some of them describing the same RNA with different names, and some of them describing, no doubt, slightly different ones with the same name. As those links above show, there are a lot of ways to analyze and detect these things, and that scrambles the interpretation pretty well, too. From what I can see, it is truly a wild and lawless frontier.

Then you read a paper like this new one. The authors selected 25 lncRNAs in zebrafish, trying to pick ones that would be as functional as possible (previously proposed function, conserved sequence, high expression, position next to important DNA sequences, etc.) Then they used CRISPR to take them out. And guess what? Hardly anything happened. The resulting mutants were all viable, developed normally, and were fertile with normal offspring. A close look showed that some of the deletion mutants had slightly altered gene expression profiles for nearby sequences, and even those had no detectable phenotype.

So what does this mean? The authors say “LncRNAs might have redundant, subtle, or context-dependent roles, but extrapolation from our results suggests that the majority of individual zebrafish lncRNAs are dispensable for embryogenesis, viability and fertility“. It’s hard to argue with that, unless they managed to pick 25 loser sequences somehow. It could be that while they’re functional, they’re easily compensated for by other lncRNA species during development, which if true has implications for a lot of the hypotheses about their roles in the cell. It could be that in some disease states the roles of particular lncRNAs become much more defined – or not. As I said, the wild frontier. Welcome to it!

45 comments on “You Don’t Understand Long Noncoding RNAs”

  1. myma says:

    Ah, Biology.
    I say “Good Luck with that”.

  2. AR says:

    Biological investigations continue to uncover deeper levels of regulation and complexity, how can we possibly integrate this field into target based drug design practices. The future of screening and drug design is phenotypic and polypharmacological.

    1. RA says:

      Tired of hearing this trite mantra I’m afraid. Phenotypic screening is part of the future just like it’s part of the present, but there are almost as many assumptions and dotted lines between a phenotypic assay and a disease as there are between a target and a disease. I’ve seen them fail to translate into patients enough times to think it’s not the holy grail some believe.

      1. AR says:

        Probably more dotted lines in some cases, but still reckon it’s preferable.

  3. Kevin H says:

    So, a zebrafish is just a lncRNA’s way of making more lncRNA?

    1. Anon says:

      Actually we’re all just vehicles for making and carrying around incRNAs. That’s our real purpose in life.

  4. rcyran says:

    My pet theory, with zero evidence, is that they are a storehouse of sequences that may be potentially useful in unforeseen circumstances. I like this this idea because it’s very difficult to disprove/prove.

  5. Rule (of 5) Breaker says:

    I wonder if these are vestigial. Back in primordial days RNA did the job of proteins. Perhaps these are the last remnants of the transition to peptides as work horses. Maybe they had continued functionality even up to a few hundred million years ago. Why the conservation if they are non needed? Who knows, but if nature hasn’t had a reason to eliminate them, they made fade from evolution slowly. Let me admit that as a non-biologist I find it easy to throw out these conjectures without being weighed down by any sort of evidence for or against.

    1. John Wayne says:

      Continuing the above thought, if an organism exists under conditions wherein nucleotides are limited one might expect selective pressure to have diminished the amount of noncritical RNAs being made. Does such an organism/culture condition exist? If so, anybody know of the data?

      1. Wavefunction says:

        It’s perfectly possible that noncritical RNAs arose at some point in evolution and played a useful role, but later, when they weren’t required it was just easier for evolution to keep copying them than to get rid of them. That happens all the time and can explain junk DNA for instance; it’s just cheaper to duplicate unnecessary stuff rather than eliminate it (which would entail complicated genomic rearrangements).

        1. luysii says:

          When there is little selective pressure to keep polynucleotides around, and some to get rid of it, they can be lost quickly. The leprosy organism is relatively new, since sequence comparisons imply that it diverged from TB 14 megaYears ago. Since it lives mostly in man it has jettisoned about a quarter of its genome during this time (3,268,203 nucleotides vs. 4,441,529 for M. TB). Of the protein coding genes that are left, 1,116 are pseudogenes and 1,604 are still active. So parts of the genome can be lost ‘quickly’ when there is no reason to keep them, or if there is selective pressure to keep the genome small, as there would be in a microorganism.

          1. Stephen says:

            true but the number of breeding cycles of the leprosy organism is orders of magnitude more than humans over the same time

          2. Anonymoose says:

            Not only the number of breeding cycles, but mammals are less disposed to genomic change than bacteria on the whole. Not a fair comparison in my opinion.

      2. Toby says:

        Prokaryotes have no introns and relatively little intergenic DNA, and are under strong selective pressure to express proteins and to minimize cell division time, in potentially nucleotide-scarce conditions.

    2. Todd says:

      That makes perfect sense, considering that they’re specifically transcribed, but don’t have any specific use. Since their transcription doesn’t have much of a cost for these cells, why get rid of them in evolution?

    3. artkqtarks says:

      If this were the case, you would think that lncRNAs have ancient evolutionary origins. It seems that evolutionary ages of lincRNAs tend to be young. If I understand this paper ( right, more than 80% of human lincRNAs are primate specific.

  6. zmil says:

    “They tend to be rather cell-type specific, and even cell-cycle or developmental-stage specific, which definitely argues for their utility. ”

    This is a common misconception. While we expect transcriptional noise to show up in all sorts of cell types, there’s no reason that specific non-functional transcripts can’t be cell-type specific. Binding sites for cell-type specific transcription factors can show up in random sequences just as easily as constitutive TF binding sites. In fact, we see this with endogenous retroviruses, where very similar proviruses are expressed in very different cell types due to a few sequence differences.

  7. luysii says:

    Life is said to have originated in the RNA world. We all know about the big 3 important RNAs for the cell, mRNA, ribosomal RNA and transfer RNA. But just like the water, sewer, power and subway systems under Manhattan, there is another world down there in the cell which doesn’t much get talked about. These are RNAs, whose primary (and possibly only) function is to interact with other RNAs. They are forgotten but not gone, as was said of TB before the AIDS epidemic.

    For more about microRNAs, ceRNAs, and circular RNAs — see

    If you have a subscription, have a look at Cell vol. 174 pp. 350 – 362 ’18 for a network of regulatory RNAs acting (on each other) in the mammalian brain. It is devilishly complex.

  8. Nesprin says:

    There’s another possible interpretation- that these serve a role in extremis not seen in normal function, or that they serve as redundant backups to some other mechanism which is intact in normal zebrafish. After all, p53 loss even in humans does not preclude normal development, viability and fertility.

  9. another interesting point was how many off-target effects CRISPR produced on neighbouring genes. Funny how despite this, there was a stark absence of a phenotype. Another study did a CRISP-ability test on all lncRNAs, and apparently only 38% of them can be modified without affecting neighbouring genes due to proximal promoters/overlapping coding genes etc. Covered in this blogpost:

    1. loupgarous says:

      Isn’t that the generic issue with CRISPR deletions, that they can cause unexpected insertions and deletions?

      Whether the CRISPR targets are coding or noncoding, diligence in looking for distant off-target indels is called for. It’s not obvious to me how an off-target insertion or deletion resulting from CRISPR deletion of noncoding DNA will assuredly also not code for a phenotypic entity. What, specifically, would assure that?

      1. Hi loupgarous, I can only speak from our experience with RNAi. The probability of an off-target phenotype depends on two things – reagent specificity and assay biology. An assay measuring cell growth for example is affected by many genes. A non-specific reagent that hits multiple off-target genes is therefore more likely to impact this phenotype. An assay with a more narrow focus, dependent only on < 10 genes for example, has a lower chance of exhibiting a change in phenotype despite multiple off-targets. The probabilities were calculated by our bioinformatician in this blogpost:

        In this case, I would have thought Zebrafish development is dependent on multiple genes, but then again, have not worked with them so can't be completely sure. The morpholinos however seem to be creating phenotypes, so to some extent perhaps CRISPR is more specific than RNAi, though not completely free from off-targets.

        1. meant to say rather the CRISPR knockouts may be creating less off-targets than knock-downs morpholinos (which work not via RNAi rather but through an antisense mechanism). Alternatively, the lack of phenotypes could also be due to genetic compensation which we’ve seen happening in knock-outs.. alot

  10. artkqtarks says:

    I may be cynical, but I think many researchers want to believe lncRNAs are important because they want lncRNAs to be like splicing, catalytic RNAs, and RNAi – a new RNA-related field where they can make big discoveries.

    In 2007, Howard Chang’s lab at Stanford published a paper on a lincRNA called Hotair in Cell with a big fanfare ( The first author, John Rinn, got a faculty position at Harvard. In 2016, a paper that disputes the main conclusion of the 2007 paper was published. ( You can also read a commentary here:

    I’m not saying that lincRNAs including Hotair have no functions at all. But I do think that researchers that study them have a tendency to overhype them. They do have incentives to do so.

  11. exGlaxoid says:

    At work, we have myriads of old SOPS, manuals, notebooks, out of date references, and much more paperwork. In fact we just shipped some off for archival storage, rather than just scan it in and toss it. So I would guess that the cells must have a rather large bureaucracy of DNA in the form of admins, provosts, clerks, etc who are scared to part with anything that might be needed one day, and would rather just keep it, as opposed to just trashing it and lightening the load on all of us. Just like the government still have many useless departments, many of which do nothing tangible, but we can’t possibly cut the military wool procurement department, Dept. of civil war veterans affairs, or Spanish American War tax on phones.

  12. Charles Hixson says:

    One possibility is that they’re a waste product from some other reaction. Being long makes this a bit less plausible, but I’m not sure how much less plausible. How quickly does each one degrade?

  13. Josephine says:

    Most lncRNAs are most likely not functional, as some previous commenters have pointed out. Larry Moran at Sandwalk is a good resource for understanding the controversy better. A sample:

  14. Shanedorf says:

    This discussion reminds me of the story by Alan Bellows: On the Origin of Circuits where they tried to mimic evolution in creating better microchips. What they found was that there were some pathways that seemed useless, but turned out to be critical to function based on the electromagnetic quirks of the local environment.

  15. cynical1 says:

    I think they are there for our alien forefathers to use when they arrive and take over our bodies and turn us all into H.R. Giger type monsters. But by then, we’ll have CRISPR all worked out and we’ll be able to silence their RNA and save our planet from the big bad aliens. Someone better call Signourney Weaver for the next installment……..

  16. ScientistSailro says:

    Off topic, but I noticed today that there are two commentators who are probably women, based on their chosen name. As a long-time reader, that seems unusual. Derek, do you have any information on the gender distribution of you readership or commentariat?

    1. Derek Lowe says:

      None at all, actually – and no way of finding out for sure, either, as far as I know. . .

      1. ScientistSailor says:

        I bet Google could figure it out…

  17. t says:

    Interesting…if there doesn’t appear to be anything special about the sequence, could these just serve to change the nature of the solute environment in the cell…or act to deaggregate other cellular components?

  18. Barry says:

    In the 25million years since sea lions diverged from other carnivora (bears, cats, dogs, mustelids) they’ve ditched hundreds of olfactory GPCRs. But I don’t know if those genes are wholly deleted, or (more likely) just broken

  19. Insilicoconsulting says:

    Is it remotely possible that it may have something to do with infectious biology? Vestigial remnants or not, omehow related to viral infections?

  20. Jb says:

    People always assume everything is biology based. What if lncrnas are made by cells for a physicochemcial reason? To alter the viscosity of the cytoplasm or something?

    1. Derek Lowe says:

      That’s not at all a crazy idea, and it’s strengthened by the current vogue for phase-transition condensate studies in cells. That’s been proposed as a reason for “excess” ATP levels in the cytoplasm, for example.

  21. LBC says:

    This whole debate becomes even weirder when you consider the fact that most promoters, at least in humans, are bidirectionally transcribed. See e.g. this paper

    1. Derek Lowe says:

      I still find that an unnerving result. My mental pictures of how transcription works are totally inadequate, and I suspect that I have a lot of company there.

  22. Anon says:

    Some kind of insurance policy for the cells? We all think that paying insurance is big waste of money and such, that is until something happens. A kind of comfort. Same with the cells, i suppose.

  23. DrOcto says:

    Just guessing here, but could they just be for nucleotide storage?

  24. maw says:

    So viable, developed normally, and were fertile with normal offspring. Doesn’t sound very exhaustive to me – just what they could easily measure.

    Reminded again of ‘could a neuroscientist understand a microprocessor’

  25. Emjeff says:

    They’ve got to have a connection to Alzheimer’s. Quick, give me some VC money!

  26. oligoman says:

    Aren’t a lot of these lncRNA’s the result of retroviruses?

  27. X says:

    If I had to guess – and boy howdy, do I – I would point to the tendency of evolved systems to use big, complex regulatory systems bursting with positive and negative feedback loops, and suggest that the lncRNA serves some sort of regulatory or protective function that’s opaque to the usual “let’s pull out one piece and see what happens” experimental model.

    But I also like the “extra nucleotide storage” and “cytoplasm viscosity” ideas…

Comments are closed.