Tuesday, October 21, 2025

Google AI references a "Biblical Genetics" video in claiming that junk DNA is no longer considered junk

Today I did a routine search for "junk DNA" "2025" to see if misinformation is still dominating the web. It is, but that's not the most surprising thing I discovered. Here's what Google AI told me at the top of the search page.

In 2025, "junk DNA" is no longer considered junk, as new studies show it plays vital roles in gene regulation and development. Research from 2025 indicates that these sequences, many of which come from ancient viruses, can act as "genetic switches" that influence how genes are turned on or off and how cells respond to their environment. This has led to potential breakthroughs in regenerative medicine and cancer treatment by providing new therapeutic targets.

This video explains how what was once considered junk DNA has been found to contain thousands of new genes:

The video is by Robert Carter who has a Ph.D. in molecular biology. His site is called Biblical Genetics. He also posts on creation.com

Carter sounds like he knows what he's talking about but he's just parroting all the misinformation that permeates the scientific literature. The main message of this video is that scientists were shocked to discover that the human genome only had 20,000 protein coding genes but we now know (no, we don't) that each gene makes many different proteins and that accounts for the "missing" complexity that all the experts had expected.1

We also "know" (no, we don't) that scientists have discovered tens of thousands of new protein coding genes that make small proteins. He references a Science article by Elizabeth Penissi who has been spreading misinformation about the human genome for more than 25 years.

It's not surprising that Robert Carter wants to discredit the idea of junk DNA. What's surprising is that Google AI is directing readers to a creationist video.


1. The knowledgeable experts predicted that the human genome would have fewer than 30,000 genes and that's exactly what was found when the human genome sequence was published.

17 comments:

  1. Not that it would be any better to cite one of Nature's or ENCODE's absurd propaganda videos in this respect. They contain the same stupid misinformation.

    ReplyDelete
  2. It's not surprising at all that AI references a creationist video. Unfortunate, yes, but not surprising.

    ReplyDelete
  3. The answers of the Google search AI are based on the first page or so, of ordinary search results. The AI (Gemini) is more skeptical when you ask it something directly.

    ReplyDelete
  4. Are they not confusing the "no junk" fallacy with the "alternative splicing" fallacy? If we hypothetically found lots of functional isoforms of known proteins that wouldn't reduce the amount of junk DNA by one bit.

    ReplyDelete
  5. Sorry, I seem to default to "anonymous" now; that's new.

    ReplyDelete
  6. Pervasive functional translation of non canonical human open reading frames
    Jin Chen https://orcid.org/0000-0002-6634-4397, Andreas-David Brunner

    It's time to pay more attention on non_canocial ORF which have discovered to have important roles in immunity and development . They're beyond noisy transcript

    ReplyDelete
  7. @Anonymous: Please describe the evidence that a substantial number of these ORFs encode a biologically functional protein/peptide. Some of that evidence might include showing that they are conserved or that they are subject to purifying selection.

    Please estimate the total amount of functional DNA that these ORFs would add to the genome if every single one of them were functional. I'm guessing that it's about 0.1% or less.

    ReplyDelete
    Replies
    1. The paper I cited, just investigated a hundred number of non-concial ORF . By the way, I assume your guess about the proportion of junk ones, is purely based on sequence conservation. I'm afraid but that method of detecting functionally important sequence through sequence conservation has not a good reputation. Since many non conserved regions of the genes found to be quite essential and vise versa

      Delete
  8. Latest, most direct test of the neutral theory soundly disproves it again.

    José Aguilar-Rodríguez
    @jaguilarrod
    One of the most exciting projects of my career, years in the making. Using high-throughput precision genome editing, we mapped the fitness effects of thousands of natural variants—challenging the idea that common variants are inconsequential.
    https://www.biorxiv.org/content/10.1101/2024.10.30.621178v2 Massively parallel interrogation of the fitness of natural variants in ancient signaling pathways reveals pervasive local adaptation

    ReplyDelete
    Replies
    1. Note that these are amino acid changes in proteins. Nothing to do with junk DNA. Also note that they found most variants to be neutral. Direct quote from the abstract: "We address this question by mapping the fitness effects of over 9,000 natural variants in the Ras/PKA and TOR/Sch9 pathways—key regulators of cell proliferation in eukaryotes—across four conditions in Saccharomyces cerevisiae. While most variants are neutral in our assay, ∼3,500 exhibited significant fitness effects."

      Delete
  9. The Kimura neutral theory of 1968 was inspired by protein-sequence-alignment-based genetic distances among species (DNA sequence came much later), such as the genetic equidistance phenomenon (GEP). The GEP was surprising and inspired the molecular clock hypothesis, which directly inspired the neutral theory. NT considers most observed variants or amino acid substitutions as neutral. When applied in population genetics, NT treats most coding and non-coding DNA variants as neutral. So, NT considers both coding and non-coding variants as neutral, regardless of whether between species or within species.

    The study identified a substantial proportion of the examined genetic variants as non-neutral, contrary to predictions from the neutral theory (NT). Why not all or most variants? This is likely attributable to the limited scope of their functional assays, which evaluated only four specific conditions. In nature, organisms face far more diverse and stringent challenges—environmental stresses, pathogens, metabolic demands, and ecological interactions—that extend well beyond these four criteria. Consequently, more comprehensive and exhaustive testing would be expected to reveal functionality in the majority of variants.

    ReplyDelete
    Replies
    1. The study identified a substantial proportion of the examined genetic variants as non-neutral, contrary to predictions from the neutral theory (NT).
      Note that you have confused the prediction of the neutral theory (most variation neutral) with a strawman (all variation neutral), and have gone from the previous universe of study (coding and non-coding) to a highly restricted one (coding, non-silent). Note that even in the study you cite, 2/3 of the variants are evolving neutrally.

      Delete
  10. @Anonymous says, "The paper I cited, just investigated a hundred number of non-concial ORF."

    The paper identified some transcripts that are bound to ribosomes. They looked at the ones producing a detectable peptide. They provided some evidence that these may be functional. The total number is 230 uORFs and 91 from lncRNAs.

    " By the way, I assume your guess about the proportion of junk ones, is purely based on sequence conservation."

    "The average size of these micropeptides is less than 100aa so even if we assume they are all derived from real genes (not junk) this only amounts to a tiny fraction of functional DNA (<0.1%). I'm anxiously awaiting confirmation of the data implying function.

    "I'm afraid but that method of detecting functionally important sequence through sequence conservation has not a good reputation. Since many non conserved regions of the genes found to be quite essential and vise versa."

    I'm aware of spacer DNA sequences that show conservation of size but not nucleotide sequence. What percentage of the genome do you think is functional but shows no evidence of conservation or purifying selection? The most recent data on the amount under purifying selection is about 10% and that's similar to the amount that is conserved.

    If you are going to challenge the concept of 90% junk then you better have a good case for arguing that a substantial amount of intron sequences and transposon debris actually has a function in spite of the fact that it is evolving neutrally.

    I look forward to hearing you make that case.

    ReplyDelete
    Replies
    1. I confidently assume that our genome is entirely functional because of ever increasing data in reasrch articles discovering new function for transposable elements introns , psydogenes and even repetitive satellites. (No matter how many article I cite, you always keep saying that these are only the minority ones ).

      Delete
    2. I confidently assume that our genome is entirely functional because of ever increasing data in reasrch articles discovering new function for transposable elements introns , psydogenes and even repetitive satellites. (No matter how many article I cite, you always keep saying that these are only the minority ones ).

      Delete
  11. The study concludes “What is surprising and inconsistent with nearly neutral theory is that many very large-effect polymorphisms are present at high frequencies in the population. These results provide strong support in favor of the balance theory of natural variation.” https://www.biorxiv.org/content/10.1101/2024.10.30.621178v2

    The so-called balance theory of natural variation is in fact encompassed by the Maximum Genetic Diversity (MGD) theory, which posits that genetic diversity (or distance) is maintained at an equilibrium corresponding to a maximum or optimal level. In principle, genetic diversity must either have such a maximum equilibrium or not. Empirical evidence that refutes the latter view — as assumed by the neutral theory — necessarily supports the former, thereby vindicating the MGD framework.

    Approximately 44% (1,612/3,629) of non-synonymous variants exhibit significant non-neutral effects, compared to ~33% (1,134/3,473) of synonymous variants and ~37% (866/2,349) of noncoding variants. These findings reveal only modest quantitative differences in non-neutrality across variant types, rather than the qualitative distinctions predicted by the neutral theory, which assumes most variants are neutral and unaffected by selection, especially for the syn and noncoding variants.

    In short, most of the so called evidence for the neutral theory are based on unproven or unrealistic assumptions and none includes empirical test data. All empirical tests have invalidated it and vindicated its only possible rival.

    ReplyDelete