science – bioliteracy / biofundamentals

W. F. R. Weldon’s Critique of Mendel’s Work: Biological Relevance and Context.

A RAG-ChatGPT written backgrounder (checked and edited by Mike Klymkowsky ) for the excessively curious – in support of biofundamentals – 1 August 2025

Weldon’s Background and Perspective: Walter F. R. Weldon (1860–1906) was a British zoologist and a pioneer of biometry – the statistical study of biological variation. He believed that evolution operated through numerous small, continuous variations rather than abrupt, either-or traits. In his studies of creatures like shrimps and crabs, Weldon found that even traits which appeared dimorphic at first could grade into one another when large enough samples were measured [link]. He and his colleague Karl Pearson (1857-1936) argued that Darwin’s theory of natural selection was best tested with quantitative methods: “the questions raised by the Darwinian hypothesis are purely statistical, and the statistical method is the only one at present obvious by which that hypothesis can be experimentally checked” [link]. This emphasis on gradual variation and statistical analysis set Weldon at odds with the emerging Mendelian school of genetics, led by William Bateson (1861-1926) that focused on discrete traits and sudden changes. By 1902, the scientific community had split into two camps – the biometricians (Weldon and Pearson in London) versus the Mendelians (Bateson and allies in Cambridge) – reflecting deep disagreements over the nature of heredity [link]. This was the charged backdrop against which Weldon evaluated Gregor Mendel’s pea-breeding experiments.

Critique of Mendel’s Pea Traits and Categories Weldon’s photographic plate of peas illustrating continuous variation in seed color. (This figure from his 1902 paper shows pea seeds ranging from green to yellow in a smooth gradient, contradicting the clear-cut “green vs. yellow” categories assumed by Mendel [link]. Images 1–6 and 7–12 (top rows) display the range of cotyledon colors in two different pea varieties after the seed coats were removed [link]. Instead of all seeds being simply green or yellow, Weldon documented many intermediate shades. He even found seeds whose two cotyledons (halves) differed in color, underscoring that Mendel’s binary categories were oversimplifications of a more complex reality [link].

Weldon closely re-examined the seven pea traits Mendel had chosen (such as seed color and seed shape) and argued that Mendel’s tidy classifications did not reflect biological reality in peas. In Mendel’s account, peas were either “green” or “yellow” and produced either “round, smooth” or “wrinkled” seeds, with nothing in between. Weldon showed this was an artifact of Mendel’s experimental design. He gathered peas from diverse sources and found continuous variation rather than strict binary types. For example, a supposedly pure “round-seeded” variety produced seeds with varying degrees of roundness and wrinkling [link]. Likewise, seeds that would be classified as “green” or “yellow” in Mendel’s scheme actually exhibited a spectrum of color tones from deep green through greenish-yellow to bright yellow [link]. Weldon’s observations were impossible to reconcile with a simple either/or trait definition [link].

Weldon concluded that Mendel had deliberately picked atypical pea strains with stark, discontinuous traits, and that Mendel’s category labels (e.g. “green vs. yellow” seeds) obscured the true, much more variable nature of those characters [link]. In Weldon’s view, the neat ratios Mendel obtained were only achievable because Mendel worked with artificially generated lines of peas, bred to eliminate intermediate forms [link]. In ordinary pea populations that a farmer or naturalist might encounter, such clear-cut divisions virtually disappeared: “Many races of peas are exceedingly variable, both in colour and in shape,” Weldon noted, “so that both the category ‘round and smooth’ and the category ‘wrinkled and irregular’ include a considerable range of varieties.” [link] In short, he felt Mendel’s chosen traits were too simple and unrepresentative. The crisp binary traits in Mendel’s experiments were the exception, not the rule, in nature. Weldon’s extensive survey of pea varieties led him to believe that Mendel’s results “had no validity beyond the artificially “purified”in-bred” races Mendel worked with,” because the binary categories “obscured a far more variable reality.”[link]

Mendel’s Conclusions and Real-World Heredity. Weldon went beyond critiquing Mendel’s choice of traits – he questioned whether Mendel’s conclusions about heredity were biologically meaningful for understanding inheritance in real populations. Based on his empirical findings and evolutionary perspective, Weldon doubted that Mendel’s laws could serve as general laws of heredity. Some of his major biological objections were:

– Traits are seldom purely binary in nature: Outside the monk’s garden, most characteristics do not sort into a few discrete classes. Instead, they form continuous gradations. Weldon realized that Mendel’s insistence on traits segregating neatly into “either/or” categories “simply wasn’t true,” even for peas [link]. Mendel’s clear ratios were achieved by excluding the normal range of variation; in the wild, peas varied continuously from yellow to green with every shade in between [link]. What Mendel presented as unitary “characters” were, in Weldon’s eyes, extremes picked from a continuum.

– Mendel’s results were an artifact of pure-breeding: Weldon argued that the famous 3:1 ratios and other patterns were only apparent because Mendel had used highly inbred, “pure” varieties. By extensive inbreeding and selection, Mendel stripped away intermediate variants [link]. The artificially uniform parent strains used in Mendel’s experiments do not reflect natural populations. Weldon concluded that the seeming universality of Mendel’s laws was misleading – they described those special pea strains, not peas (or other organisms) at large [link]. In a letter, he even mused whether Mendel’s remarkably clean data were “too good” to be true, hinting that real-world data would rarely align so perfectly [link].

– Dominance is not an absolute property: A cornerstone of Mendelism was that one trait form is dominant over the other (e.g. yellow dominates green). Weldon questioned this simplistic view. He gathered evidence that whether a given trait appears dominant or recessive can depend on context – on the plant’s overall genetic background and environmental conditions [link]. For example, a seed color might behave as dominant in one cross but not in another, if other genetic factors differ. Weldon argued that Mendel’s concept of dominance was “oversimplified” because it treated dominance as inherent to a trait, independent of development or ancestry [link]. In reality (as Weldon emphasized), “the effect of the same bit of chromosome can be different depending on the hereditary background and the wider environmental conditions”, so an inherited character’s expression isn’t fixed as purely dominant or recessive [link]. This questioned the biological generality of Mendel’s one-size-fits-all dominance rule.

– Atavism and ancestral influence: Perhaps most intriguing was Weldon’s concern with reversion (atavism) – cases where an offspring exhibits a trait of a distant ancestor that had seemingly disappeared in intervening generations. Breeders of plants and animals had long reported that occasionally a “throwback” individual would appear, showing an old parental form or color after many generations of absence. To Weldon, such phenomena implied that heredity isn’t solely about the immediate parents’ genes, but can be influenced by more remote ancestral contributions [link]. “Mendel treats such characters as if the condition in two given parents determined the condition in all their offspring,” Weldon wrote, but breeders know that “the condition of an organism does not as a rule depend upon [any one pair of ancestors] alone, but in varying degrees upon the condition of all its ancestors in every past generation” [link]. In other words, the influence of a trait could accumulate or skip generations. This idea directly conflicted with Mendel’s theory as presented in 1900, which only considered inheritance from the two parents and had no mechanism for latent ancestral traits resurfacing after several generations. Weldon concluded from examples of reversion that Mendel’s framework was biologically incomplete – there had to be “more going on” in heredity than Mendel’s laws acknowledged [link].

In sum, Weldon found Mendel’s laws too limited and idealized to account for the messy realities of inheritance in natural populations. Mendel had demonstrated elegant numerical ratios with a few pea characters, but Weldon did not believe those results scaled up to the complex heredity of most traits or species. Variation, continuity, and context were central in Weldon’s view of biology, whereas Mendel’s work (as interpreted by Mendel’s supporters) seemed to ignore those factors. Thus, Weldon saw Mendel’s conclusions as at best a special case – interesting, but not the whole story of heredity in the real world [link][link].

Weldon’s Legacy

Weldon’s critiques came at a time of intense debate between the “Mendelians” and the “Biometricians.” William Bateson, the chief Mendelian, vehemently defended Mendel’s theory against Weldon’s attacks. In 1902, Bateson published a lengthy rebuttal titled Mendel’s Principles of Heredity: A Defense, including a 100-page polemic aimed squarely at “defending Mendel from Professor Weldon”[link]. Bateson and his allies believed Weldon had misinterpreted Mendel and that discrete Mendelian factors really were the key to heredity. The clash between Weldon and Bateson grew increasingly personal and public. By 1904 the feud had become so heated that the editor of Nature refused to publish any further exchanges between the two sides [link]. At a 1904 British Association meeting, a debate between Bateson and Weldon on evolution and heredity became a shouting match, emblematic of how divisive the issue had become [link][link].

Although Weldon’s objections were rooted in biological observations, many contemporaries saw the dispute as one of old guard vs. new ideas. Tragically, Weldon died in 1906 at the age of 46, with a major manuscript on inheritance still unfinished [link]. In that unpublished work, he had gathered experimental data to support a more nuanced theory reconciling heredity with development and ancestral effects [link][link]. With his early death, much of Weldon’s larger critique faded from the spotlight. Mendelian genetics, championed by Bateson and later enriched by the chromosome theory, surged ahead. Nevertheless, in hindsight many of Weldon’s points were remarkably prescient. His insistence on looking at population-level variation and the importance of multiple factors and environment foreshadowed the modern understanding that Mendelian genes can interact in complex ways (for example, polygenic inheritance and gene-by-environment effects). As one historian noted, Weldon’s critiques of Mendelian principles were “100 years ahead of his time” [link]. In the context of his era, Weldon doubted the biological relevance of Mendel’s peas for the broader canvas of life – and while Mendel’s laws did prove fundamental, Weldon was correct that real-world heredity is more intricate than simple pea traits. His challenge to Mendelism ultimately pushed geneticists to grapple with continuous variation and population dynamics, helping lay the groundwork for the synthesis of Mendelian genetics with biometry in the decades after his death[link][link].

Sources: Weldon’s 1902 paper in Biometrika and historical analyses [link][link][link][link][link][link][link]provide the basis for the above summary. These document Weldon’s arguments that Mendel’s pea traits were overly simplistic and his laws of heredity not universally applicable to natural populations, especially in light of continuous variation, context-dependent trait expression, and atavistic reversions. The debate between Weldon and the Mendelians is detailed in contemporary accounts and later historical reviews [link][link], illustrating the scientific and conceptual rift that formed around Mendel’s rediscovered work.

Introductory genetics: one way by which determinism creeps into biology students’ heads.

background: I have long been interested in students’ (and the public’s) misconceptions about biology (see this & that). More and more, it appears to me that part of the problem arises when conventional biology (and science courses in general) leave underlying scientific principles unrecognized and/or unexplained. In biology, there is a understandable temptation to present processes in simple unambiguous ways, often by ignoring the intrinsic complexity and underlying molecular scale of these systems. The result is widespread confusion among the public, a confusion often exploited by various social “influencers”, some (rather depressingly) currently in positions of power within the US.

After attending a recent Ray Troll and Kirk Johnson roadshow on fossils, art, and public engagement at the Denver Museum of Nature and Science (DMNS), I got to thinking. As a new hobby, in advance of retirement, perhaps I can work on evolving the tone of my writing to become less “academical” and more impactful, engaging, and entertaining (at least to some) while staying scientifically accurate and comprehensible. So here goes an attempt (helped out by genAI).

A common misconception, promoted by some “science popularizers” is that biological systems, including humans, are “determined” or “super-determined” (what ever that means) by various factors, particularly by the versions of genes, known as alleles, they inherited from their parents. While there is no question that biological systems are influenced and constrained by a number of factors (critical to “stay’n alive), the idea of determinism seems problematic (considered here). So where would a belief in biological determinism come from? One possibility, that emerged in the “Teaching and Learning Biology” course taught with Will Lindsay@CU Boulder, is the way basic genetics is often presented to students. The specific topic that caught my attention was the way the outcome of genetic crosses (matings) was presented, specifically through the use of what are known as Punnett’s squares.

In a typical sexually reproducing organism, the parents with different mating types or sexes, e.g. male and female, have two copies of each gene (mostly) – they are termed “diploid”. The two versions of a particular gene can be the same, in which case the organism is said to be “homozygous” or different, when it is termed “heterozygous” for that gene. The allele(s) carried by one parent can be the same or different from the allele(s) carried by the other. Molecular analysis of the alleles present in a population has been key to determining who, back in the day, was mating with Neanderthals (see wikipedia). Each gamete (egg or sperm) produced contains one or the other version of each gene – they are termed “haploid”. When sperm and egg fuse, a new diploid organism is generated.

Much of what is described above was figured out by Gregor Mendel (wikipedia). The good monk employed a few tricks that enabled him to recognize (deduce) key genetic “rules”. First, he worked with peas, Pisum sativum and related species. He used plants grown by commercial plant breeders to have specific versions of a particular trait. In his studies, he focussed on plants that displayed versions of traits that were unambiguously distinguishable from one another. Such pairs of traits are termed dichotomous; they exist in one or the other unambiguously recognizable form, without overlapping intermediates. The majority of traits are continuous rather than dichotomous.

As part of the process of generating “predictable plants”, breeders select male and female plants with the traits that they seek and discard others. After many generations the result are plants with reproducible and predictable traits. Does this mean that the plants are identical? Nope! There is still variability between individual plants of the same “strain”. For example, Mendel used strains of “tall” and “short” pea plants; the tall plants had stem lengths of between ~6 to 7 feet while the short plants had stem lengths between ~0.75 to 1.5 feet (a two-fold variation)(see Curtis, 2023). He put them into tall and short classes, ignoring these differences. But these plants were different. Such differences arise through stochastic processes and responses to developmental and environmental effects that impact height in various ways (discussed in a past post). Mendel began his studies with 22 strains of pea plants but only 7 exhibited the dichotomous behaviors he wanted. If he had included the others, it is likely he would have been confused and never would have arrived at his clean genetic rules. In fact, after he published his studies on peas, he took the advice of Carl Nägeli (see wikipedia) and began studies using Hieracium (hawkweeds), which differs in its reproductive strategy from Pisum (Nogler 2006). Nägeli’s suggestion and Mendel studies lead to uncertainty about the universality of Mendel’s rules. Mendel’s experiences reflect a key feature of scientific studies: simplify, get interpretable data, and then extend observations / systems leading to confirmation or revision.

The variation inherent in biological systems is nicely illustrated by what is (or should be) a classic study by Vogt et al (2008) who described the variations that occur within populations of genetically identical shrimp raised in identical conditions. The variation between genetically similar organisms (or identical twins) found in the wild (natural populations) is much greater. Why? because in breeder supplied plants most of the allelic variation present in the wild population is lost, discarded in the process of selecting and breeding organisms for specific traits. We see these “genetic background” effects when looking at genetically determined traits in humans as well. Consider cystic fibrous, a human genetic disease associated with the inheritance of altered versions of the CFTR gene (more on cystic fibrosis). People who inherit two disease-associated alleles of the CFTR gene develop cystic fibrosis, but as noted by Corvol et al (2015) “patients who have the same variants in CFTR exhibit substantial variation in the severity of lung disease” and this variation is associated (explained by) genetic background effects, together with stochastic effects and their developmental and environmental histories. In any of a number of studies, whenever populations of organisms are analyzed based on their genotype (which alleles they carry) the result is inevitably a distribution of responses, even when the average responses are different (for a good example see Löscher 2024).

In the case of the traits Mendel studied, he concluded that the trait was determined by the presence of different versions (alleles) of a genetic “factor”, that each organism contained two alleles, that these alleles could be the same (homozygous) or different (heterozygous), and that one allele was “dominant” to the other (“recessive”). If the dominant allele were present, it would determine the form of the trait observed. Only if both versions of the alleles present were recessive would the organism display the associated trait. The other rule was that all of the gametes produced by homozygous organism carried the same “trait-producing” allele, while heterozygous organisms produced gametes containing one or the other allele.

In 1905 Reginald Punnett introduced a way of thinking about Mendel’s matings, a diagram now known as a Punnett’s square (see wikipedia). In this figure (left below) ↓ the outcome of a mating between a male homozygous for a dominant allele and a female homozygous for a recessive allele is illustrated. All of the offspring will be heterozygous, but it is worth keeping in mind, however, that does not mean that they will be identical – they will display a similar level of variation in the trait seen in populations of the parental plants (see above). Again, this variation arises from the impact of environmental effects on developmental processes together with the influence of stochastic effects. The variation associated with the particular set of alleles present in an organism is captured by what is known as variable penetrance and expressivity of a gene-influenced trait (see link for molecular details). Ignoring the variations observed between organisms carrying the same allele(s) of a gene (or the same genotype in identical twins or clones can encourage or reinforce the idea that the details of an organism (its phenotype) are determined by the alleles it carries.

Another way students’ belief in genetic determinism can be reinforced is perhaps unintentional. Typically the result of the original mating between homozygous recessive and dominant parents (the P generation) is termed the first filial or F1 generation. Often the next type of genetic cross presented to students involves crossing male and female F1 individuals to produce the second filial or F2 generation (see figure – right above ↑). Such as F1 cross is predicted to produce organisms that display the dominant to recessive trait in a ratio of 3 to 1. What is often missing is that reproducible observation of this ratio requires that large numbers of F2 organisms are examined. The result of any particular F1 (heterozygous) cross is unpredictable; it can vary anywhere from 0 to 4 dominant to recessive trait displaying organisms to 4 to 0 trait dominant to recessive trait displaying organisms, and anything in between. This behavior is characteristic of a stochastic process; predictable when large numbers of events are considered and unpredictable when small numbers of events are considered. Stochastic behaviors are common in biological organisms, given the small numbers of particular molecules, and specifically particular genes, they contain (a GoldLabSymposium talk on the topic). In the context of organisms, there is room for something like “free will” (consideredhere). Whether Elon “knows” he is giving something that closely resembles a Nazi salute or not, we can presume that he is, at least partially, responsible for his actions and by implication their ramifications.

Why are the results of a mating stochastic? Because which gamete contains which trait-associated allele occurs by chance, while which gametes fuse together to produce the embryo is again a chance event. Some analyses of the numbers Mendel originally reported led to suggestions that his numbers were “too good”, and the perhaps he fudged them (for a good summary see Radick 2022). The bottom line – subsequent studies have repeatedly confirmed Mendel’s conclusions with the important caution that the link between genotype and phenotype is typically complex and does not obey strictly deterministic rules.

Nota bene: This is not mean to be a lesson in genetics; if interested in going deeper I would recommend you read Jamieson & Radick (2013) and the genetics section of biofundamentals.

Literature cited:

Corvol et al., (2015). Genome-wide association meta-analysis identifies five modifier loci of lung disease severity in cystic fibrosis. Nature communications, 6, 8382.

Curtis (2023). Mendel did not study common, naturally occurring phenotypes. Journal of Genetics, 102(2), 48.

Jamieson & Radick (2013). Putting Mendel in his place: How curriculum reform in genetics and counterfactual history of science can work together. In The philosophy of biology: A companion for educators (pp. 577-595). Dordrecht: Springer Netherlands.

Löscher (2024). Of Mice and Men: The Inter-individual Variability of the Brain’s Response to Drugs. Eneuro, 11(2).

Nogler (2006). The lesser-known Mendel: his experiments on Hieracium. Genetics, 172(1), 1-6.

Radick (2022). Mendel the fraud? A social history of truth in genetics. Studies in History and Philosophy of Science, 93, 39-46.

van Heyningen (2024). Stochasticity in genetics and gene regulation. Philosophical Transactions of the Royal Society B, 379(1900), 20230476.

Vogt et al., (2008). Production of different phenotypes from the same genotype in the same environment by developmental variation. Journal of Experimental Biology, 211, 510-523.

Unexpected molecular sexual dimorphisms (in host mitochondrial-microbial interactions).

Unexpected molecular sexual dimorphisms (in host-microbial interactions).

While wandering through the literature, I found myself reading a paper by departmental colleagues Dong Tian, Mingxue Cui, and Min Han (2024). They describe effects of bacterial mucopolysaccharides, components of bacterial cell walls, on mitochondrial functions in mice and human cells. While an interesting example of the interaction between components of the microbiome and its host, what was particularly surprising to me was their observation that these mucopolysaccharides had sex-specific effects on mitochondrial functions (footnote 1). This sexual dimorphism was documented by examining the generation of reactive oxygen species (ROS) formed during reactions involving molecular oxygen (O₂), concentrations of ATP (cells’ primary energy currency), and body weight, when comparing control mice with mice treated with “an antibiotic cocktail to deplete the intestinal microbiota in order to eliminate the source of muropeptides following a well-established protocol for antibiotic-induced microbiome depletion (AIMD)”. One interesting result was that the addition of bacteria-derived proteoglycan derived “muropeptides” (see their figure 3 ↓ (modified) and the included image, derived from it with scale modifications) dramatically inhibited the effects of AIMD treatment on a number of cellular behaviors in mouse and human cells.

Perhaps It should be expected that removing an organism’s internal microbiome causes a number of stress effects on the host, nor that mitochondria (derived, via endosymbiosis and subsequent evolutionary adaptations) respond to such changes. Mitochondrial functions seem particularly sensitive to general cellular stresses (a topic considered in more detail in the appearance of mitochondrial stress effects associated with “knock-out” mutations in a number of intermediate filament proteins (the work of others reviewed here).

What was surprising (certainly to me) was the observation that male and female cells responded differently to muropeptides; something reported previously by Gabanyi et al (2022). There are, of course, a number of possible and plausible reasons for such differences. For one, a recent study of the cellular (gene-protein) network involved in male gamete formation further revealed its evolutionarily ancient origin and its complexity, involving the “expression of approximately 10,000 protein-coding genes, a third of which define a genetic scaffold of deeply conserved genes” (Brattig-Correia et al., 2024). While this study focussed on the male germ lines, the tissue that generates sperm, it is likely that differences in gene expression occur throughout the body. As an example, based on analyses of protein expression in post-mortem human brain tissues Wingo et al (2023) reported that “Among the 1,239 proteins, 51% had higher expression in females and 49% had higher expression in males”.

In this light, It is tempting to speculate that the effects of mucoproteins might well extend beyond the mouse and lead to sex-specific differences in humans as well. The source of these differences could be the result of cellular and tissue specific differences sex-influenced variations in gene expression, protein activity, and cellular organization, including social interactions between cells in tissues and organs. They may arise as indirect (and complex) effects of androgen or estrogen-based hormone responses. It is interesting how these differences may or may not impact a range of physiological effects.

Footnote

Mitochondria are evolutionary derivatives of a bacterial endosymbiont(s) found in eukaryotes (organisms like us). For some background see Mitochondrial activity, embryogenesis, and the dialogue between the big and little brains of the cell.

literature cited:

Brattig-Correia et al., (2024). The conserved genetic program of male germ cells uncovers ancient regulators of human spermatogenesis. eLife, 13, RP95774.

Gabanyi et al.'(2022). Bacterial sensing via neuronal Nod2 regulates appetite and body temperature. Science 376, eabj3986. .

Tian, D., Cui, M., & Han, M. (2024). Bacterial muropeptides promote OXPHOS and suppress mitochondrial stress in mammals. Cell reports, 43(4).

Wingo et al., (2023). Sex differences in brain protein expression and disease. Nature medicine, 29, 2224-2232.

Molecular bumper cars (RNA polymerase-ribosomal interactions): their (unexpected) functional effects and how to control them

Cells are extremely complex.¹ Much of their “core” complexity appears to have been present in their last (universal) common ancestor, known as LUCA. We find it in the “conserved” molecular mechanisms and machines active in modern cells. LUCA and its offspring are membrane-bounded, non-equilibrium systems that import free energy and export entropy to maintain and repair themselves, to grow, behave, and reproduce (and all the other things living things do). One problem, however, with LUCA is that it makes speculation on the steps that preceded it impossible to know with certainty. Not withstanding claims of breakthroughs (e.g. ‘Monumental’ experiment suggests how life on Earth may have started“), it is likely that we will never know the actual steps involved; after all, the origin of life occurred billions of years ago and under rather different conditions than exist today.

Living systems “work” based on inherited, pre-existing molecular machines and mechanisms (1). The actions of these machines are fueled through coupling to thermodynamically favorable reactions taking place under non-equilibrium conditions, i.e. the living state. Looking at the details of these interactions reveals interesting and unexpected behaviors. Unfortunately, the “simple” physical-chemical underpinnings of these processes, key to understanding them, are not always presented to students effectively (2). At the same time, the complexity of cellular systems means that in practice, the link between “simple” molecular mechanisms and the behavior of a biological systems can be obscure (see 3). That said, key insights are illuminated when molecular mechanisms are examined, as illustrated by Wee et al., (2023)(4).

Emerging from LUCA, biological populations have diverged into distinct “prokaryotic” lineages: the bacteria and archaea.² Both are defined by a protein-lipid boundary layer, the plasma membrane. Within this membrane is a single internal compartment, the cytoplasm. Information is stored in cells in two forms, first in the on-going LUCA-derived living system and the second, information in the sequence of double-stranded DNA molecules. These two types of information are interdependent: the information in DNA makes sense only within a cell and the on-going cellular processes depend upon and utilize the information in DNA. In bacteria and archaea, these are circular double-stranded DNA molecules. Here we restrict our discussion to the common unicellular bacterium Escherichia coli (E. coli), one of the workhorse systems that led to an understanding of core molecular mechanisms.

E. coli hasa single circular genomic DNA molecule of ~5 million nucleotide base pairs in length; it contains about 5000 distinct genes that encode polypeptides and functional “non-coding” RNA molecules (if you are interested in numbers, check out bionumbers). An E. coli cell is rod-shaped and ~1 micrometer (10^-6 meters) long. Its genomic DNA molecule is ~1000 times longer than the cell that contains it, and a rapidly dividing cell can contain multiple copies of the genome. Genes typically contain two distinct functional regions. Regulatory regions interact with various proteins that determine whether a gene is “expressed” or not. Coding regions specify what is expressed. The first step is the synthesis of an RNA molecule; such a molecule can encode a polypeptide or a non-coding RNA.³ Non-coding RNAs can have structural, catalytic, or regulatory functions.

The first step in gene expression in all cell types is the binding of proteins to a gene’s regulatory sequences. Typically a complex of proteins leads to the binding and activation of a DNA-dependent, RNA polymerase. The RNA polymerase complex unwinds a specific region of the DNA and uses the complementary nature of nucleotide base pairing: A binding to T in DNA and U in RNA, and C binding to G in both, to synthesize an RNA molecule based on DNA sequence. Synthesis of RNAs that encode polypeptides, known as messenger RNAs (mRNAs) starts with the 5′ end of the molecule and moves toward the 3′ end (replaced ↓ soon).

In prokaryotic cells, both DNA and mRNA synthesis reactions occur in the cytoplasm. A ribosome, a molecular machine composed of multiple proteins and RNAs, can engage the 5′ end region of an mRNA as soon as it appears – before the synthesis of the mRNA is complete. The cytoplasm of a cell contains lots of ribosomes; in E. coli there are ~70,000 ribosomes per cell (more or less). This leads to some interesting and functionally significant interactions. One thing to consider, not always stressed, is that these synthetic processes are not error proof. DNA replication (DNA-directed, DNA synthesis), transcription (DNA-directed, RNA synthesis), and polypeptide synthesis (RNA-directed, polypeptide synthesis) all have an error rate, typically 1 error per ~10⁶ addition events for DNA replication and transcription. Errors can lead to mutations in the DNA, RNAs that encode abnormal proteins, or abnormal and potentially toxic polypeptides.

To deal with physical realities, these synthetic processes employ various “error correction” strategies. In the case of DNA and RNA synthesis, the polymerases involved have what is known as “proof-reading” activities. If the incorrect nucleotide is inserted into a growing DNA or RNA chain, it can be recognized; the polymerase can then “reverse” (move backward along the DNA), remove the mistakenly inserted nucleotide, and then move forward again, adding the correct nucleotide. Key here is that the polymerase is moving back and forth along the DNA strand. The result of proof-reading is to reduce the error rate of DNA-dependent DNA and RNA synthesis substantially, down to 10^-8 to 10^-10 per base pair in the case of DNA synthesis.

In the case of the RNA polymerase, the newly synthesized RNA can fold back on itself, forming what is known as a “hairpin”. This hairpin “can stabilize an elemental pause (in RNA synthesis) an allosteric interaction with the β-flap tip helix of RNAP”. What Wee et al (4) report is as the mRNA-associated ribosome moves along the RNA it unfold the hairpin and “bumps” into the polymerase, inhibiting this “pause” which increases the rate of mRNA synthesis and inhibits the polymerase’s error correction function. The resulting mRNA population has more frequent base pair changes, errors that can influence the polypeptides synthesized. While cells of all types have various “chaperone” systems that can deal with misfolded proteins that arise in response to various stresses or errors, these can be overwhelmed. The resulting misfolded (damaged) proteins can lead to cellular defects and long term effects on viability (discussed in 5).

About 1.5 billion years later (give or take), a new type of cell appeared, the result (apparently) of a symbiotic interaction between an archaeal-like “host” and a O₂-utilizing bacterium. This synthetic organism, the progenitor of the eukaryotes, differed from either type of prokaryote in that it sequestered its genome, now composed of linear DNA molecules, within a double membrane bounded “nuclear” compartment. In this hybrid cell type, DNA and RNA synthesis was confined to the nucleus, while ribosomes and polypeptide synthesis were confined to the cytoplasm. Eukaryotic cells are typically much larger that prokaryotic cells, reproduce more slowly, and are more complex in terms of the numbers of genes, and the amount of genomic DNA they contain. It is tempting to speculated that while rapidly dividing, relatively simple prokaryotic cells may be able to tolerate more mistakes in terms of the synthesis of their polypeptides, larger, more complex eukaryotic cells would be vulnerable. A plausible result would be a selection pressure to separating RNA from polypeptide synthesis.

literature cited

Alberts, B. (1998). The cell as a collection of protein machines: preparing the next generation of molecular biologists. Cell, 92, 291-294.
de Lorenzo, V., 2024. The principle of uncertainty in biology: Will machine learning/artificial intelligence lead to the end of mechanistic studies?. Plos Biology, 22, p.e3002495.
Klymkowsky, M.W., 2021. Making mechanistic sense: are we teaching students what they need to know? Developmental Biology, 476, pp.308-313.
Wee et al., 2023. A trailing ribosome speeds up RNA polymerase at the expense of transcript fidelity via force and allostery. Cell, 186, pp.1244-1262.
Klymkowsky, M.W., 2019. Filaments and phenotypes: cellular roles and orphan effects associated with mutations in cytoplasmic intermediate filament proteins. F1000Research, 8.

Footnotes

if you want brush up on you molecular biology, check out chapter 7 of biofundamentals ↩︎
Image from Govindjee – doi:10.3389/fpls.2011.00028, CC BY 3.0. Given the diversity of biological systems, these are general descriptions – often there a exceptions, but recognizing them all makes generating a coherent narrative difficult (and beyond me). Mea culpa. ↩︎
bioliteracy link: When is a gene product a protein when is it a polypeptide? ↩︎

Is it possible to teach evolutionary biology “sensitively”?

Michael Reiss, a professor of science education at University College London and an Anglican Priest, suggests that “we need to rethink the way we teach evolution” largely because conventional approaches can be unduly confrontational and “force religious children to choose between their faith and evolution” or to result in students who”refuse to engage with a lesson.” He suggests that a better strategy would be akin to those use to teach a range of “sensitive” subjects “such as sex, pornography, ethnicity, religion, death studies, terrorism, and others” and could “help some students to consider evolution as a possibility who would otherwise not do so.” [link to his original essay and a previous post on teaching evolution: Go ahead and teach the controversy].

There is no doubt that an effective teacher attempts to present materials sensitively; it is the rare person who will listen to someone who “teaches” ideas in a hostile, alienating, or condescending manner. That said, it can be difficult to avoid the disturbing implications of scientific ideas, implications that can be a barrier to their acceptance. The scientific conclusion that males and females are different but basically the same can upset people on various sides of the theo-political spectrum.

In point of fact an effective teacher, a teacher who encourages students to question their long held, or perhaps better put, familial or community beliefs, can cause serious social push-back – Trouble with a capital T. It is difficult to imagine a more effective teacher than Socrates (~470-399 BCE). Socrates “was found guilty of ‘impiety’ and ‘corrupting the young’, sentenced to death” in part because he was an effective teacher (see Socrates was guilty as charged). In a religious and political context, challenging accepted Truths (again with a capital T) can be a crime. In Socrates’ case”Athenians probably genuinely felt that undesirables in their midst had offended Zeus and his fellow deities,” and that, “Socrates, an unconventional thinker who questioned the legitimacy and authority of many of the accepted gods, fitted that bill.”

So we need to ask of scientists and science instructors, does the presentation of a scientific, that is, a naturalistic and non-supernatural, perspective in and of itself represent an insensitivity to those with a super-natural belief system. Here it is worth noting a point made by the philosopher John Gray, that such systems extend beyond those based on a belief in god(s); they include those who believe, with apocalyptic certainty, in any of a number of Truths, ranging from the triumph of a master race, the forced sterilization of the unfit, the dictatorship of the proletariat, to history’s end in a glorious capitalist and technological utopia. Is a science or science instruction that is “sensitive” to, that is, uncritical of or upsetting to those who hold such beliefs, possible?

My original impression is that one’s answer to this question is likely to be determined by whether one considers science a path to Truth, with a purposeful capital T, or rather that the goal of scientists is to build a working understanding of the world around and within us. Working scientists, and particularly biologists who must daily confront the implications of apparently un-intelligent designed organisms (due to ways evolution works) are well aware that absolute certainty is counterproductive. Nevertheless, the proven explanatory and technological power of the scientific enterprise cannot help but reinforce the strong impression that there is some deep link between scientific ideas and the way the world really works. And while some scientists have advocated unscientific speculations (think multiverses and cosmic consciousness), the truth, with a small t, of scientific thinking is all around us.

Photograph of the Milky Way by Tim Carl photography, used by permission

A science-based appreciation of the unimaginable size and age of the universe, taken together with compelling evidence for the relatively recent appearance of humans (Homo sapiens from their metazoan, vertebrate, tetrapod, mammalian, and primate ancestors) cannot help but impact our thinking as to our significance in the grand scheme of things (assuming that there is such a, possibly ineffable, plan)(1). The demonstrably random processes of mutation and the generally ruthless logic by which organisms survive, reproduce, and evolve, can lead even the most optimistic to question whether existence has any real meaning.

Consider, as an example, the potential implications of the progress being made in terms of computer-based artificial intelligence, together with advances in our understanding of the molecular and cellular connection networks that underlie human consciousness and self-consciousness. It is a small step to conclude, implicitly or explicitly, that humans (and all other organisms with a nervous system) are “just” wet machines that can (and perhaps should) be controlled and manipulated. The premise, the “self-evident truth”, that humans should be valued in and of themselves, and that their rights should be respected (2) is eroded by the ability of machines to perform what were previously thought to be exclusively human behaviors.

Humans and their societies have, after all, been around for only a few tens of thousands of years. During this time, human social organizations have passed from small wandering bands influenced by evolutionary kin and group selection processes to produce various social systems, ranging from more or less functional democracies, pseudo-democracies (including our own growing plutocracy), dictatorships, some religion-based, and totalitarian police states. Whether humans have a long term future (compared to the millions of years that dinosaurs dominated life on Earth) remains to be seen – although we can be reasonably sure that the Earth, and many of its non-human inhabitants, will continue to exist and evolve for millions to billions of years, at least until the Sun explodes.

So how do we teach scientific conclusions and their empirical foundations, which combine to argue that science represents how the world really works, without upsetting the most religiously and politically fanatical among us? Those who most vehemently reject scientific thinking because they are the most threatened by its apparently unavoidable implications. The answer is open to debate, but to my mind it involves teaching students (and encouraging the public) to distinguish empirically-based, and so inherently limited observations and the logical, coherent, and testable scientific models they give rise to from unquestionable TRUTH- and revelation-based belief systems. Perhaps we need to focus explicitly on the value of science rather than its “Truth”. To reinforce what science is ultimately for; what justifies society’s support for it, namely to help reduce human suffering and (where it makes sense) to enhance the human experience, goals anchored in the perhaps logically unjustifiable, but nevertheless essential acceptance of the inherent value of each person.

Apologies to “Good Omens”
For example, “We hold these truths to be self-evident, that all men are created equal, that they are endowed by their creator with certain unalienable rights, that among these are life, liberty and the pursuit of happiness.”