CharonY

I think it is actually fragmenting again. What for a while was considered bioinformatics have in part been reclaimed by biostatisticians, a part has peeled off to computational biology and yet another part falls under the broader umbrella of data sciences. And I think especially the latter is bound to specialize again (it seems to come and go in waves).

This is probably not quite what OP was thinking about, but in fast-moving sciences, often new terms are coined and changed in a relative fast manner. For example, the term "genomics" was coined in the 1980s and was used to describe the complete set of genes (genome) and later for the other biomolecules in a cell (e.g. proteome for proteins, transcriptome for RNA and so on). Then at some point folks like the term so much, that it started to be used in somewhat different contexts. For example, rather than using "proteome" to refer to a complete analysis of all proteins, it was eventually also used if one simply looked at more than one or two proteins. Another fancy term that has been circulating since the mid 2000s is "synthetic biology", which in many areas is now replacing the older term of molecular cloning (or molecular genetic) techniques. An interesting aspect of it, is that this rapid divergence of terminology actually seems to create a divide in literature. I found that many students only use the newer terms and thereby overlook older papers.

I thin it should be added that there is (AFAIK) no gold standard with regard to these tests. While NMR does work at least as well as the other methods, it is not clear which one is the best. And best is defined here as yielding a measurement that is clinically predictive. It is also important to note that there is more work looking into LDL subclasses as the category is rather overly broad and fluent. And there is at least some suggestion that certain smaller types of LDL might be more diagnostic, but the measurements are even trickier, as higher specificity is needed. That is actually also rather complicated. Early on, there were already suggestions that the link between saturated fat and LDL cholesterol (or specifically LDL/HDL cholesterol ratio) is not straightforward. Some studies for example suggest that reducing saturated fat in the diet hat less effect than exchanging saturated with unsaturated fat, especially cis-PUFA. There is certainly a connection somewhere but as metabolism goes, it is likely again indirect. Some data suggests that the issue with saturated fat is perhaps not really only in the realm of LDL/HDL cholesterol ratio, but perhaps by increasing the very small LDL cholesterol particles, which might be more likely to cause cardiovascular events. Edit: I have not read the paper referenced above, but the graph shows a slight increase in LDL-cholesterol but a shift of the most problematic (small) species from 45.2 nmol/dL to 39.6. It is unclear to me whether those changes are actually clinically relevant, but I suppose that would require a study on its own.

Nutritional guidelines are necessarily simplified and there will be a lot of individual variance. Generally speaking, once on is a healthy place, it is often helpful to check from deviations a given homeostatic situation, but even that is an imperfect approach. Sometimes recommendations are actually correct, but for incorrect reasons.

The details are actually quite complicated, but I can start simple (and my apologies if it is too trivial) and maybe start with comments on some of the trickier parts. As you mentioned, the terms LDL and HDL do not refer to the cholesterol itself. Rather, cholesterol is transported packaged by lipoproteins, the mentioned high-density lipoproteins (HDL) and low-density lipoproteins (LDL). In addition there are also very low density lipoproteins (VLDL) and intermediate density lipoproteins (IDL). The measurements therefore refer to the fraction of cholesterol associated with particles of specific density that circulate in the bloodstream. To complicate matters on this level a bit, there are slightly different assays that measure the fraction of LDL in different ways (often indirectly, e.g. using the Friedewald equation, whereas direct methods often also measured IDL and VLDL). There is some data suggesting that using ApoB (which only not associated with HDL) could be a better biomarker for cardiovascular health, but that is under discussion, too. But one way I think about VLDL- IDL-LDL is that they are different maturation steps where the very large VLDL are reduced in size and then can enter the intima. Now, originally it was believed that LDL is a transport vehicle to move cholesterol to peripheral tissue and organs and HDL moves surplus cholesterol back to the liver. In part, the idea is then that very high LDL-cholesterol leads to deposits that can cause arteriosclerosis, for example. However, when trying to look at associated mechanisms, things get complicated pretty fast. For example, it was found that the vast majority of cells actually have an active lipid metabolism and most cholesterol are produced where they are used and are not necessarily delivered via LDL. Then, there is the issue that a lot of LDL cholesterol is derived from HDL and a lot of them is taken up by LDL-receptors in the liver. I.e. of the LDL is actually directed to, not away from the liver, making it questionable whether delivery to the periphery is really the main function of LDL. Likewise, HDL has been known to be critical for cholesterol efflux capacity (removing cholesterol from macrophages and transport to liver), but now studies suggest that LDL amplifies these efforts by HDL pathways. So taken together, the classic dichotomy of LDL vs HDL (cholesterol) has become rather questionable but we do not have a fully articulated model yet that can be used for better health prediction. Edit: I should add that my expertise is mostly limited to biomarker analysis, and not the clinical aspects, so it is therefore biased a bit more on the molecular/analytical side and may not reflect clinical standards. Therefore none of it should be considered medical advice of any sorts.

It is actually quite complicated. Originally there was the distinction between "bad" and "good" cholesterol (LDL vs HDL) and to some extent it still makes sense. Specifically LDL/HDL ratio seems to be somewhat diagnostic for cardiovascular risk, but there are quite a few caveats. There are some discussions on whether there is a functional role of LDL for e.g. atherosclerosis, or whether it is more of a side effect. Similarly, it was assumed that dietary intake of cholesterol would find its way into the bloodstream. However, newer findings suggest that dietary cholesterol has only an indirect impact. There is a ton of new findings, and robust discussion on that matter, but no simple answers yet. But for OP the studies suggest that cholesterol in eggs does not really have a direct impact on lipid profiles (i.e. it is not absorbed and released in the blood stream). However, high cholesterol food can apparently influence your own lipid homeostasis, but in many cases the effect was fairly moderate. I have not seen the video, but if the persons is arguing about absorption of dietary cholesterol, they are working on rather obsolete information.

Also it evaporates rather fast, making it rather useless (and dangerous) as a deterrent.

I think it is not only an issue of power (though it can be) but to a large extent it is exhaustion. The acts of what is now called microaggression often were normalized and even if pushback was successful, dealing with it on a regular basis, especially as it likely would not change anything, makes it easier to just ignore than to fight. I also suspect that this continues to be the modus operandi for most folks, but social media allows the voices that do object to be magnified (for better or worse). Context. Which is why it is mostly futile to try to come up with "objective" rules. Eh, I don't see how it has changed. Mostly about what folks get offended by, actually. Try to talk to folks about how stereotyping whole groups of people is harmful maybe one or two decades ago and see them implode because they feel that you accused them racism.

Much of it is due to the way hypotheses are set up and/or our tendency to create categories to make data and experiments easier to handle. The good news is that with better technology we also get more quantitative data. The downside is that new insights often have trouble penetrating public perception, which often prefers simple categories and narratives (for obvious reasons). I suspect it can also go the opposite way. I do not think that we can always expect that cultural norms necessarily follow the direction hinted at by biology.

I have not gone through all the posts, so I am sorry if I am being repetitive. However, it seems to me that there still quite a bit of back and forth on the nature vs nurture thing. And I think the evidence we have in general on that matter is that it is almost always not a "vs" but an "and" situation. Biological systems in general are highly dynamic and responsive to external and internal inputs and often do not behave in a strictly deterministic way (which is why I dislike the way "nature" as a term is used in this context). Even the behaviour of very simple systems, like cells can be difficult to predict unless you control virtually every environmental input. And even then we see differences within a clonal cell population. Development, especially of the brain, is highly interactive and even cues during fetal development can affect the outcome. As such, for virtually any trait we have a certain proportion of genetic vs developmental/environmental component. However, figuring the exact distribution is extremely difficult. It also does not help that there is a lot of assumptions regarding gender or sex preferences which often introduce bias in studies. Dissecting cultural effects (or impact by the test system, like the clever Hans effect) can further muddy the waters. The reproducibility crisis in psychology, which at least in part is related to a big narrative with little data issue and the fact that sometimes the studies are poorly standardized. Either way, as with most biological systems we are clearly looking at a continuum of behaviour that has to be explored quantitatively. Trying to force a clear binary system out of it tends to be problematic and is likely going to boil down into semantic exercises.

Developmental biology is not really my forte, but a lot of developmental timing is related to humoral control (which affects gene expression), which is also dependent on factors such as nutrition. So there is a bit of a complex interplay and I am actually not sure whether there is simple model explaining these differences somewhere.

With regard to differences we should differentiate aspects related to development/maturation, which does seem to have measurable sexual dimorphism, to structural aspects. The knowledge on the latter and especially the question whether sex dimorphism in human brains actually exists is an everchanging discussion, which is highly dependent on our technology to measure these differences. In mature humans, the effect size seems to diminish, the more we measure. A fairly recent meta-analysis argues that most differences described are related to brain size rather than any functional differences, for example. This is, of course, not the question in OP, but I thought it might be worthwhile to mention, in case the discussion drifts into that area.

Well, there is a bit of nuance, as mentioned earlier, one of the question in some of the standard surveys specifically asked whether folks would be uncomfortable with and showing that the comfort is highly contextual. And this is really the issue when trying to frame it as an innate response. The distaste for being approached is not only dependent on the suitor, but highly dependent on context. I would bet that in a highly professional environment most folks would feel uncomfortable when approached overtly sexually, regardless of ones sexual orientation, for example. Moreover, I think I may mentioned that folks might be put off by by sexual advances of any folks they are not interested in. While this mechanism might have some innate components, it is clear that it is not specifically targeted at homosexuality. I.e. there are many cues at play that can trigger the distaste reaction. We learn, for example, that this behaviour has no place at work. We might be more receptive in other social settings. So in short, the feeling of distaste might only be partially learned, but the cues triggering those are social and hence, learned (and therefore also malleable). This is what I tried to address with my comment about hardwired before.

To your first point, I think it the term hardwired or innate does obscure some of the mechanisms surrounding sexual orientation. Most likely it is a developmental mechanism where genetic factors contribute, but not necessarily determine sexual orientation. What we do know is that typically it is fixed at an early age. However, one should contrast them to sexual preference, which might be finer grained. I.e. the attraction among the perceived sexually compatible partners. These are much clearer to be learned, but are likely also heavily influenced by child-hood learning. There are several mechanisms described in psychology in that regard where childhood might influence partner selection. There is for example a hypothesis in psychology, called the Westermarck effect which assumes a form of inprinting in which folks tend not to be attracted to siblings, if they have lived together at a young age. It is an attractive hypothesis as in contrast to what is under discussion here, there is a path to selective advantages (i.e. avoiding incest). The problem though is if experimental data does not really support such a mechanism. What has been found is that e.g. disgust with incest is more related to social and cultural cues, though the debate is not fully settled yet. So even from a perspective where at least theoretically there could be strong selective factors and which appear to be a automatism, the underlying mechanisms are apparently far more complex. And obviously there is not really a good argument to made for strong selection on mutable traits. As with many things, I think the somewhat unsatisfactory answer is that most behaviour, even many unconscious ones are learned on one level or another. Our brain requires constant feedback to develop and some behavioural traits (such as sexual orientation) can be fixed very strongly, whereas others remain malleable. The OP was talking about selection and as such the traits that are malleable are not under selection. However, the basis for such traits (e.g. the mechanisms which influence how we develop sexual preferences) might be. And I think in the discussion so far, both factors have been mixed up.

It falls roughly under the same umbrella. There is a host of data, starting from the one I mentioned above where it shows that folks are far less uncomfortable (including sexual advances from individuals from the same sex). I think it is fairly clear that this is not true for sexual orientation. I.e. it only seems similar if you use the terms of like and dislike. But I think you will agree that sexual orientation is far more hardwired (though part is likely learned early on), whereas the other "dislike" is far more malleable.

The logic is that sexual orientation basically does not change through life, making it either genetic or at least strongly imprinted early in life and in a mostly unchangeable way. By the same token, homophobic sentiments can change quickly as noted earlier. Thus, using the same reasoning, it is much less likely to be innate.

I just want to add that group selection has been in discussion for a long time and despite some resurgences, most evolutionary scientists find it problematic, as other than just-so stories, it has not been useful in explaining persistence of traits. Moreover, many social traits can and have been explained in the context of "regular" selection more efficiently. The whole idea kind of nosedived together with the sociobiology and evolutionary psychology.

Considering how much a person has learned by that age, including one or more languages, and many other unnatural things (wearing clothes, hygiene, culture-specific mannerisms). Moreover, "natural" reactions can also have learned components. Imprinting is such an example, for example (though perhaps less well studied in humans compared to other animals).

Great, except folks won't get tenure if they do that. See the economic aspect in OP.

A degree is a low bar. Do you think folks who are unable to pass it should go into these positions? Do you really think that a job interview will screen out inadequate performance better than a few years of training and supervision?

Sounds good, if you are able to ignore about 80% of any given class. That is true, and in many disciplines writing is a big part of it. One way you seem to suggest to do writing in class only? Because that is where we are headed for the moment. I used to run a lab where the objectives where research-driven. I.e. teaching methods and then present them with small questions that they had to figure out using the methods they were taught. The idea was to get away from rote memorisation to application of knowledge. I co-ran a version as postdoc a long time ago with quite some success. A little while ago I tried it again, and basically one one student liked it (who is now a postdoc) whereas the rest specifically complained about the amount of work and the lack of simple and direct answers. In many countries the university system has slowly changed from a somewhat elitist to a much broader system. Unfortunately that has also changed attitudes of the student body. While in the past it was mostly pre-meds, now quite a few other students are also getting anxious when they do not have clear question-answer sheets that they can use to guide their learning. Whereas lectures were used in the past to augment reading, folks do not read anymore and you get massive complaints if the answer to a question is not prominently featured in one of the lectures. Mind you, my experience is not at an elite school, but some of my colleagues who teach in ivy league school see a more muted version of what we experience in lower tier schools. But so far the solution seems to be focus on in-class performance, which is basically what we are doing right now. Though assignments and homework are a bit more questionable. I am not a fan of those in the first place, but then folks complain about not having extra credits if you do not give them the opportunity to turn those in. That all being said, I think there is some worth to have nurses and physicians who are able to read and comprehend texts without assistance, I think.

Then perhaps tell me, how do you think education should work. Should we just provide content and let the students sort it out? Or hand out degrees for enrolling. I.e. are there perhaps any thoughts on the system?

So how do you educate folks who are mostly in for a degree?

Here is also a short article on the WHO comments on that matter: https://arstechnica.com/science/2023/01/covid-is-still-a-global-health-emergency-but-end-may-be-near-who-says/

Inspired by another thread, I wanted to see opinions on the rise of language AI for science teaching. I would actually like to go back a bit to the rise of the internet. Before that time, literature search was an often slow process involving going into the library and copying articles manually. With the rise of online databases and encyclopedias, access to information became a breeze. The in hindsight optimistic assumption of that time is that folks would need to spend less time searching for info, and therefore spend more time synthesizing. Thus, the logic went the quality of student work (but also research) should improve, as folks would use more time on thinking and less on rote memorization. I think most educators at this point will realize that this is not what happened. While folks memorize less, they also seem to read (and by extension) think less than they used to, paradoxically perhaps because they have access to a vast array of literature. The lack of reading is nowhere as obvious when students are writing essays. However now AI systems are making the rounds which is able to generate well-written (if frequently inaccurate) essays, articles or whatever you want. As the massive cheating during the pandemic has shown us, many, if not most students will use any means to improve their scores with as little effort as possible. Here again essays were seen as a way out as it requires more than a quick google search, compared to exam questions. But obviously ChatGPT is going to make it more difficult for the educator. So in the light of these modern developments, how should modern teaching look like? What should educators do in order to assess academic abilities? I also want to add that in many countries college-level education has strong economic incentives, where university administrations tries to get as many satisfied customers as possible, whereas increasingly students tend to focus on grades rather than improvement in their understanding (not least due the high cost). So we have an unfavorable situation of economic incentives and technological developments that, in my opinion, are negatively impacting learning and at this point I do feel that is a bit more than just a generational complaining issue. So I would like to have an open discussion on the (hopefully) various perspectives on this issue.