More Recent Comments

Saturday, March 02, 2024

Nils Walter disputes junk DNA: (4) Different views of non-functional transcripts

I'm discussing a recent paper published by Nils Walter (Walter, 2024). He is trying to explain the conflict between proponents of junk DNA and their opponents. His main focus is building a case for large numbers of non-coding genes.

This is the third post in the series. The first one outlines the issues that led to the current paper and the second one describes Walter's view of a paradigm shift. The third post describes the differing views on how to define key terms such as 'gene' and 'function.' In this post I'll describe the heart of the dispute according to Nils Walter.

-Nils Walter disputes junk DNA: (1) The surprise

-Nils Walter disputes junk DNA: (2) The paradigm shaft

-Nils Walter disputes junk DNA: (3) Defining 'gene' and 'function'

How Distinct Frameworks Engender Divergent views of ncRNAs

The title of this section is taken directly from the paper. Nils Walter describes the email exchanges he had with Alex Palazzo and me last year. He explains that our differing views on junk DNA boil down to whether we think non-coding RNAs (ncRNA) are functional or not. Here's how he opens that discussion ...

As a biophysical chemist who uses single molecule microscopy, in my respectful, if sometimes antagonistic, electronic discussions with the group of junk DNA champions I naturally adopted the viewpoint that function is a feature of the individual molecule.

He goes on to explain that when you focus on individual molecules, such as transcripts, he prefers the view that they are functional because they exist even if an exact function hasn't been identified. He doesn't express it like that but that's what he seems to be saying—the default conclusion is that transcripts are functional. In this sense, he echoes the 2012 ENCODE definition of function. But, as I pointed out in earlier posts, he is unaware of the fact that ENCODE researchers abandoned that position shortly after September 2012 and they've never defended it since.

Junk DNA proponents prefer to identify functional elements by testing to see whether they are being preserved by natural selection (purifying selection). In fact, we DEFINE function by this criterion. Walter's view of this position seems a little confusing.

To try to mitigate my own bias, I will refer to the function of an individual molecule as “elemental”, in contrast to the “phenotypical” function that is the higher bar used by geneticists and evolutionary biologists for the definition of the function of, for example, a gene. Based on this difference in definition, critics of ENCODE have pointed out that merely showing the existence of a process, such as transcription of an RNA, or an elemental (or rudimentary, minimal) biochemical activity, such as the transcribed RNA binding a protein, is not sufficient to prove phenotypic (or, in their view, functional) significance.

He views the maintenance definition of function as "phenotypical." I assume this is because it involves testing to see whether deletion of an element affects the viability of the individual. In this sense, he is correct because junk DNA proponents claim that deleting some DNA elements that are transcribed would have no effect on the survival of the individual; therefore, they are junk DNA. I suppose that's a "phenotypical" function because if you delete the sequence it doesn't affect the phenotype.

What's disappointing is that Walter doesn't explain why his opponents make such a claim. It's because the evidence for spurious transcription and inappropriate binding of proteins to nucleic acids (DNA or RNA) is conclusive. We know this happens so you can't just assume that when you see a transcript or an interaction it must be functional. We also know (now) that most of the sequences that are transcribed are evolving at the neutral rate. In other words, they are not under purifying selection.

Nils Walter views this controversy as a philosophical disagreement.

On some level, then, the ENCODE controversy is a philosophical disagreement on how much of an ncRNA's function has to be discovered before legitimately calling it functional, and not junk.

I think this is simplistic but there's a sense in which part of the dispute could be called "philosophical." That's the issue of the null hypothesis and where the burden of proof lies. I have yet to read a serious defense of the idea that the null hypothesis should be that all transcripts are functional unless you can prove that they have no function. On the other hand, there are plenty of serious arguments in favor of the idea that the null hypothesis should be no function and the burden of proof lies with those who claim function.

We intend to provoke alternative interpretation of questionable evidence and thorough inquiry into unsubstantiated claims.

Ponting & Haerty (2022)

One description of this position comes from a remarkable review by Chris Ponting and Wilfried Haerty published more than a year ago in Annual Review of Genomics and Human Genetics (Ponting and Haerty, 2022)1 [see Most lncRNAs are junk]. Here's a quote from the Ponting and Haerty paper with a reference to Ulitsky and Bartel (2013).2

The cellular transcriptional machinery does not perfectly discriminate cryptic promoters from functional gene promoters. This machinery is abundant and so can engage sites momentarily depleted of nucleosomes and rapidly initiate transcription. The chance occurrence of splice sites can then facilitate the capping, splicing, and polyadenylation of long transcripts. A very large number of such rare RNA species are detectable in RNA-sequencing experiments whose properties are virtually indistinguishable from those of bona fide lncRNAs. Consequently, “a sensible [null] hypothesis is that most of the currently annotated long (typically >200 nt) noncoding RNAs are not functional, i.e., most impart no fitness advantage, however slight” (Ulitsky and Bartel, 2013: p. 26).

Junk DNA proponents point out that for most transcripts there's no evidence of function and that's exactly what you would expect if they were junk RNA. Walter addresses one particular aspect of this point, namely the fact that deleting or mutating the DNA responsible for the transcript has no observable phenotype. Normally you would take this as evidence in support of spurious transcription but if you are looking to defend function you need to find a way around this observation.

The main argument is usually some version of "we just haven't looked hard enough for a phenotype" and that's the one that Walter replies on.

In fact, genetically deleting an ncRNA may not immediately lead to an observable phenotype. That fact, however, is typical of complex, critical biological functions that often involve subtle functional effects (such as that of individual miRNAs), cell-type- and stress-specific functions, or redundant pathway components to ensure a robust biology.

This sounds a lot like ad hoc rescue. Faced with the obvious conclusion; namely, that deleting a segment of DNA has no effect, therfore it is junk, you try to escape the obvious by imagining a way to preserve your belief. There certainly are examples of DNA sequences with subtle functional effects that we could cherry-pick from the literature but whether they can be generalized to several hundred thousand ncRNA genes is a different story. The observation is that spurious transcripts exist, AND many of these nRNAs are not conserved AND they are not under purifying selection AND they can be deleted without noticeable effect. You may be able to come up with a few examples of functional non-coding genes that might possibly have these properties but it seems like extreme hand-waving to argue that they are abundant.

I don't think this is a philosophical disagreemnt. It's a dispute over how scientists interpret data and when you should abandon a failed hypothesis.

1. It's interesting that this paper isn't mentioned in the Walter essay even though it's one of the most devastating critiques of his position.

2. The Ulitsky and Bartel (2013) paper is also not mentioned in Walter's essay. Perhaps that's because Ulitsky and Bartel are biochemists working on RNA and they say, "A sensible hypothesis is that most of the currently annotated long (typically >200 nt) noncoding RNAs are not functional, i.e. most impart no fitness advantage, however slight. Like all biochemical processes, the transcription machinery is not perfect and can produce spurious transcripts that have no purpose."

Ponting, C.P. and Haerty, W. (2022) Genome-Wide Analysis of Human Long Noncoding RNAs: A Provocative Review. Annual review of genomics and human genetics 23. [doi: 10.1146/annurev-genom-112921-123710]

Ulitsky, I. and Bartel, D.P. (2013) lincRNAs: genomics, evolution, and mechanisms. Cell 154:26-46. [doi: 10.1016/j.cell.2013.06.020]

Walter, N.G. (2024) Are non‐protein coding RNAs junk or treasure? An attempt to explain and reconcile opposing viewpoints of whether the human genome is mostly transcribed into non‐functional or functional RNAs. BioEssays:2300201. [doi: 10.1002/bies.202300201]


Anonymous said...

"I don't think this is a philosophical disagreemnt. It's a dispute over how scientists interpret data and when you should abandon a failed hypothesis."

Put that way, I'd see it as a philosophical disagreement lying in the subfields of epistemology and philosophy of science. But it my be more productive to examine the ontological assumptions that underlie the positions.

Larry Moran said...

@Anonymous: Do you know any philosophers who are willing to defend Nils Walter’s views on this issue?

John Harshman said...

Will you be discussing his views on the inadequacy of sequence non-conservation and genome size variation as indicators of junk? Because it's not even clear to me what he thinks his argument is.

Graham Jones said...
This comment has been removed by the author.
Graham Jones said...

I don't understand his arguments either. In particular, this paragraph leaves me completely baffled.

"Notably, however, the human body contains ~30 trillion cells, of which ~300 billion
are destroyed daily since they have stopped functioning properly.[95]
Combined with the observed genomic mosaicism of our body’s cells,
enhanced by ncRNAs, these observations imply that the sequenced
genome of an individual human being itself is the outcome of relentless
internal evolutionary pressures to conform to the needs of all bod-
ily functions. Such considerations are absent from arguments invoking
the inability of natural, purifying selection to weed out mildly dele-
terious mutations among the relatively small size (compared to the
human genome) of the human population, which regards only the entire
organism as the target of evolution, therefore likely inaccurately sug-
gesting that neutral evolution and drift give rise to the many genomic

He's not talking about cells in the germline, so what's the relevance?

Anonymous said...

@anonymous, a scientific disagreement over interpretation of data, within the confines of science, is not philosophical epistemology.

Anonymous said...

My reply is meant for anon @3:54.

SPARC said...

@Graham Jones
I don't understand that either. I guess it may have something to do with the fact that he comes from the Eigen lab. Thus, this may be a reference to Manfred Eigen's hyper-cycles.
I had three colleauges who were working in Eigen's lab and I never really understood them. I had the impression they rather dismissed straight forward explanations and preferred more complicated ones that they unfortunately could not express in compenesable terms. Of course the problem might rather have been on my side.

Joe Felsenstein said...

The problem with the designation of "without noticeable effect" is that the ability of a model organism to survive and reproduce in a lab is not a rigorous test of lack of effect. Remember that the "drift barrier" is having a selection coefficient greater than 1/N. For a natural population of modest size, say 1,000,000, drift dominates if s < 0.000001. I doubt that most laboratory observations would even be able to detect a selection coefficient of s = 0.001. The effect that is noticeable by natural selection over millions of generations, leading to conservation between species, is a lot smaller than the effect that is noticeable by us in the lab.

Which is not to be interpreted as me agreeing with Walter's other arguments.

Joe Felsenstein said...

Correction of my comment: drift dominates when the absolute value of s is less than 0.000001.

Larry Moran said...

@Joe Felsenstein: I realize now that by saying "without noticeable effect" I opened myself up to the criticism that you expressed.

There are two separate issues here. The first concerns the definition of function. Do we described it in terms of whether or not a given sequence contributes to the survivability of the individual or species? Or do we use the causal role definition preferred by Walter and other junk DNA opponents?

The second issue is how do we test the maintenance function definition. That's the part that you challenge and I agree that it's practically impossible to do an experiment that would definitely answer the question. Deleting a few megabases of presumably junk DNA should actually have a tiny positive effect on fitness but we'll never be able to see it.

Larry Moran said...

@Joe Felsenstein: I should clarify that when I said that deleting some presumably junk DNA should have a slightly positive effect that would only be true in a very large panmictic population.

Suppose in some individual we deleted 5 Mb of DNA that was truly junk and s = 0.000001. Now let's say we take that individual and 999,999 randomly selected humans and send them all off to another planet. We'll pass a law saying you can't mate with anyone who is closer to you than 5th cousin.

Wait 30 million years (one million generations).

The probability that the deletion will become fixed is approximately 2s, or 0.002%, right?

What's the probability if we delete the same 5 Mb in 500,000 people before sending them all off to another planet? I'm sure you can do that calculation. I can't because I'm not good with logarithms.

Joe Felsenstein said...

@Larry: A very practical experiment. Elon Musk will presumably be advocating it any day now.

Jathro said...

Would not Walter's definition also include every metabolic disease and even death - say via cancer? I mean all of that is biological 'function.' The maintenance seems more rigorous.

Larry Moran said...

@Jathro: Junk DNA opponents often point to the fact that mutations in non-coding DNA are associated with genetic diseases. They do this is order to refute the strawman version of function that restricts function to coding DNA.

We've always known that mutations in regulatory sequences can cause problems so this isn't news. There's plenty of room for functional non-coding DNA in the 10% of the genome that's functional.

However, there's another problem with that argument. It's possible for mutations in junk DNA to cause genetic diseases. Known examples include mutations in the junk DNA part of introns that create new aberrant splice sites.

Larry Moran said...

@Mahrshad asks, " you think science need philosophy? why not/ why yes?"

Let's be clear about one thing. Philosophy needs science.

I don't know if the reverse is true. I know that all disciplines need to follow the standard rules of logic and argumentation and that's clearly the domain of philosophy. So, in that sense, everyone needs philosophy.

I'm still looking for good examples of philosophy making a significant contribution to science where the contribution lies outside of science itself and clearly within the domain of philosophy. That's going to be a challenge since I adopt the broad definition of science that encompasses philosophy.