The Science Age

For example, it doesn't actually seem that hard to begin life from a biochemical standpoint. The particulars of how our branch of life began are difficult to ascertain and we haven't synthesized any totally artificial species, so that's not to say it's easy, but given the ubiquity of certain elements and exponential ways or ordering them, as well as the relatively young age of Earth, it's not theoretically unlikely. Of course there may be limits we haven't considered yet, and some scientists do put stricter boundaries on the possible biochemical bases of life, but it's a reasonable and inspiring idea.

On the other hand, evolution isn't directional, it simply favors species that can stay alive. Since staying alive doesn't require complex intelligence, it's completely possible (even likely, if you consider humans to be the only intelligent species on Earth) that alien life would not have it. Moreover, even if an alien species was intelligent, it could be a very different kind of intelligence. Different environmental conditions and evolutionary pressures could emphasize completely different processes - for instance the hypothetical presented by a SETI researcher in this article, where smell, rather than sight or sound, is the primary sense, precluding the development of abstract mathematics because they are not particularly apparent or useful.

As noted in the project's description, looking at the solar system in this way is somewhat of an exercise in tedium. Planets are massively far apart, on a scale that's hard to really visualize. The act of scrolling across so much empty space helps to demonstrate that, as does the corollary that it's impossible to show the solar system at both accurate sizes and distances on a regular computer screen. If you shrank everything down to fit within 1280 pixels, nothing, not even the sun, would be big enough to fill one pixel. (And that's just calculating the distance from the sun to Neptune; express your love for Pluto all you want, it's not a planet.) It really helps put into perspective why things like interplanetary missions, let alone the Voyagers, take years or decades to complete. With that in mind, I'd really recommend scrolling rather than clicking through the illustration (there are informative and funny comments in space along the way to keep you occupied).

The deeper under water you go (as with any fluid on earth) the more pressure that water exerts on you. As the paper notes, descending 100 meters (the length of a football field) increases the pressure by 10 atm (for reference, 1 atm is the about the pressure we experience here on the surface of the earth). Seeing as the deepest parts of the ocean are around 11,000 meters deep, that’s an enormous amount of pressure. Living things are sensitive to pressure on a number of structural scales, from organs (which can rupture at excessively high or low pressures) to individual cells (likewise, killing them and impairing tissue functions) to proteins (which can deform, rendering them biologically useless).

To deal with issues of pressure on protein deformation, fish have evolved a molecule, called TMAO (trimethylamine N-oxide) which stabilizes proteins within their cells. TMAO counteracts the effects of water pushing into proteins at high water pressures, allowing the proteins, and therefore the fish, to survive in deep water. 

At the same time, fish cells (like all vertebrate cells) are sensitive to how many particles are within them relative to how much water. Those particles (like proteins and nutrients) are critical for a cell’s life, but so is the amount of water (to maintain its structure and biochemical processes). The ratio of the two, the concentration inside a cell, must be maintained on a very particular range or the cell will die. Vertebrates depend on having a concentration within their cells that is a certain ratio lower than the concentration outside their cells. For fish, that outside is the ocean.

Based on some earlier research, the researchers had reason to believe that the amount of TMAO necessary to keep proteins functional is proportional to the amount of pressure: the greater the pressure, the greater the amount of TMAO must be present in cells. Fish, therefore, have two competing necessities. On one hand, they need TMAO at increasing concentrations to survive at lower depths. On the other, they need their overall cellular concentrations to be lower than that of seawater. The researchers of the PNAS paper calculated the theoretical limit imposed by this balance: at 8,200 meters below the surface of the ocean, a fish would need so much TMAO that it’s cellular concentration would match that of the surrounding water. So, fish should not be able to survive below 8,200 meters, since having the same or a higher cellular concentration would kill it.

To test this idea, the researchers measured the concentration of TMAO in related fish species living at different depths, from 900 to 7,000 meters. They found that there really was a proportional relationship between the depth and the concentration of TMAO in the fish. They also found that those concentrations lined up with their predictions, supporting the idea that 8,200 meters is an approximate biochemical limit. These findings, paired with the fact that no living fish has ever been found below 8,370 meters (with that one record-holder standing far ahead of the rest of the pack, which have been above 7,700 meters), support the proposal of a depth limit for fish based on a biochemical balance between TMAO and ocean salinity.

The authors acknowledge that their evidence doesn’t prove this to be the case. Just because no fish have been found beyond that limit so far doesn’t mean there aren’t any, or there are other factors related to depth (such as nutrient availablity) that could be the real cause of the limit. Alternately, the cause could be related to TMAO but not ocean salinity: it could simply be that the concentrations of TMAO required for greater depths are toxic to the fish. Or, everything proposed in this paper might be correct, but there may be deeper-living fish that have evolved an independent mechanism to resolve its pressure and concentration problems. Those gaps being pointed out, it's a fascinating finding in a emerging area of research. 

The idea of directionality seems to be a by-product of the fact that humans defined evolution. We see ourselves as bigger, more complex, more social, and more intelligent than the other things living today and assume that this is better. We assume that evolution has built up to us, the ideal living things. That is just not true. It may be the case that we are bigger or more intelligent (although this is evaluated by our own tests, so it's tricky to know what we're really measuring), but there are obviously smaller and less intelligent things still surviving. There is no reason, from the perspective of evolution, that one subset of currently living things would be ideal and none of the myriad of others, or that one set of traits would be more ideal than the others.

A recent article in Nautilus covers a lot of this issue in more detail (although it implies a lot more contemporary sympathy for directionality than I've encountered in the scientific community). The popular support of directionality certainly does originate in the (antiquated) scientific acceptance of it though. Fortunately, evolutionary biologists have started to correct the directionality bias, visibly evident in the revised, non-hierarchical "tree" of life diagrams.

Moreover, directionality implies intentionality. If evolution is working towards something, it has to know where it's going. This blatantly contradicts the mechanisms of evolution, which are random alterations played out in survival and reproduction. Evolution is a convenient name for a natural process (similar, in that way, to gravity), it's not a conscious entity with control over living things.

The idea of directionality in evolution, in addition to just being wrong, leads to insidious misunderstandings, such as the popular idea that humans have stopped evolving or that other living things are less evolved than us. If you look at the actual concept of evolution, this idea is absurd. Nothing living can ever stop evolving. The rate of mutations can change, as can the particular selective pressures, but evolution itself never stops. Either a group is alive and evolving, or extinct.

Scientific American has an interesting article describing the idea of "rainbow gravity." Essentially, rainbow gravity says that, different wavelengths of light react differently to the same gravitational influence. The technical aspects are beyond me, but according to Scientific American this idea could account for the separate worlds of general relativity (which as we understand it today operates on large scales but not tiny ones) and quantum mechanics (which covers the tiny scales but currently does not work at large ones). Additionally, it has interesting implications for the origin of the universe. Whereas most models start from the familiar big bang, rainbow theory does not. There is currently no evidence to support rainbow gravity over other ideas, but it is a possibility, and experiments have been proposed that are beginning to be technologically feasible which could bolster or detract from it.

Nautilus has a great article following the work of John Wheeler, who (among many other things) proposed an alternate way of unifying general relativity and quantum mechanics. Wheeler is famous for conducting the delayed choice extension of the double slit experiment. Within the rules of quantum mechanics, the behavior of light counterintuitively depends on whether it is being observed or not. The double slit experiment shows that a unit of light can either take one of the possible paths available to it (like a person taking one route to the grocery store, which makes intuitive sense) or all possible paths at once (like one person simultaneously walking every possible way to the store, which is very weird) depending on whether you monitor which path it takes. Wheeler took this a step further by realizing that, in quantum mechanics, it doesn't matter when you make this observation. The delayed choice experiment shows that you can actually causally determine how the light behaves (taking one path or all of them) by choosing to observe the path or not after the light has already started traveling. This apparent reversal of cause and effect is mind-boggling, but can be understood and has been proven to happen in quantum mechanics. Wheeler proposed that the entire universe might be a quantum system, and that our act of observing the big bang could actually have caused it. A new version of this idea is supported by Steven Hawking and Thomas Hertog in "top-down cosmology," which they aim to test using measurements of the Cosmic Microwave Background.

While the stages to come will pose their own challenges, the ships rotation was one of the greatest uncertainties. Salvage planners had to balance the need to learn as much as possible about the structural state of the ship with the fact that every hour it spends in its current state furthers its decay. In an engineering operation of this size, it simply isn’t possible to know all the variables ahead of time; planners were making their safest, best, educated guess. And it wasn’t without its risks. Rotating the ship caused even more strain on its structure, in places and directions it wasn’t designed to withstand. It’s already been under a huge amount of stress lying on its side for the past year and a half. If the ship wasn’t quite sound enough in any one place during the rotation it could break apart, littering the coral reef with debris and ruining any chance of a complete salvage. On the other hand, it was possible for the supports salvagers put in place to facilitate the rotation to break under the strain, rolling the whole ship down the rocky hill it’s been precariously resting on.

The good news is that salvagers have dodged the first problem. The Concordia successfully rotated (over the course of almost a full day) without any extra damage to the hull, allowing it to come fully upright. Now it’s resting on the underwater platform, awaiting installation of additional ballast chambers that will be used to float the remains of the ship as high above the water as possible, preparing it for the tow away. Not that these steps will be much easier. As this Scientific American piece notes, the sponsons for the damaged side need to be individually formed and attached, and the process isn't slated to be done until the spring. And then, of course, final adjustments have to be made for the refloating and towing. I'll continue posting updates as the operation progresses!

You can see a time-lapse video of the rotation maneuvers at Time. 

The process is massive in scale but relatively easy to understand. Salvagers have already wrapped the ship in chains attached to anchoring stanchions to hold it in place. No one wants it to shift while the rest of the project is being set up. The rest of the project involves building a scaffold platform on the seabed that will be able to hold the ship up after it’s been tipped over, and attaching cassions, giant ballast containers that can alternately be filled with water or air to adjust the ship’s buoyancy, to the side of the ship that tipped above the sea’s surface. When everything is in place, cables will slowly pull the above-water side down towards the platform, while the chains will help it twist back into its normal position. Once there, more cassions will be attached to the formerly below-water side, and together all of the cassions will slowly be filled with air, lifting the wrecked ship to its regular floating height. Then other boats can tow the bandaged wreck away from the island to shore.

In one recent study researchers used eyetracking during short video clips to differentiate between people with different mental disorders or healthy controls. That makes it really interesting as a potential diagnostic aid; the process, one set up, is entirely automatic, and it can give a diagnosis to a specified degree of uncertainty in 15 minutes. It’s important to note that in this study, eyetracking was just being used as an independent factor. Researchers feed the data they’ve gathered into an algorithm that picks up on patterns in the groups they define. After it learns those patterns well enough, it can tell you which group it thinks a new person is in based on their eyetracking data. It does this exceptionally well, but not necessarily in a way that means anything for the diagnosis. Altered eye movement patterns are not inherently a symptom of Parkinson’s or ADHD, they are just convenient data that happen to vary depending on what diagnosis a person has (although researchers are planning on seeing if they can tie eye movements to a source in actual cognitive symptoms). That doesn’t mean eyetracking isn’t extremely useful or clinical practical, it just means you need to be careful in interpreting it.

These types of proxy measures are also useful for studying non-human organisms, which don’t have the confounding factor of consciousness, but also can’t speak to tell us what we’re interested in. Eyetracking was recently been used to further clarify what peahens (female peacocks) look for in choosing a mate. Previous research had used behavioral measures on different interventions; cutting off some feathers’ eyes, for example, meant that a peacock was less favored. Eyetracking is a more direct way of determining which qualities are important. Since peacocks attract their mates visually, it makes sense that the most direct research would examine what the peahens are actually looking at. In this case, the researchers found that peahens spend most of their time looking at peacocks to the width of their tail fan, not individual feather's eyes. This isn't a result they could have predicted from the behavioral information; eyetracking opened up a new line of questions that wouldn't have otherwise been revealed.

UBiome’s lawyers advised that research needs IRB approval if it: receives federal funding, is a clinical trial, or will be published in a peer-reviewed journal. This covers pretty much all “professional” science. Much of the research in the US today is funded by government institutions like NIH or NSF, because they’re some of the only organizations with the resources & inclination to cover the costs of contemporary scientific research. The specific clause for clinical research brings pharmaceutical & interested biotech companies under the umbrella of requirements. Finally, peer-reviewed journals are one of the best, most respected ways of sharing your results with the rest of the world. If you’re a scientist looking to do something with your research, you’re going to (try to) publish it in a peer-reviewed journal, which gives it an audience and stamps it with a seal of legitimacy. (These things vary from journal to journal, but “peer-review” denotes that the research article submitted is critiqued by other researchers familiar with its area, so research that is not well-designed or satisfactorily analyzed won’t be accepted. While it has its weak points, it essentially serves to weed out bad science. It’s one of the most successful cornerstones of scientific research.) These categories intentionally cover a wide swath to capture any human research that then could have a potential for abuse.

If those requirements are met, there are different “levels” of IRB requirements depending on what the research actually proposes to do. It’s a very technical process designed to define as explicitly as possible what constitutes concepts like “research” and “involving human subjects” which seem clear enough until you try to tease apart which studies have the potential to abuse participants and which don’t. All in all, they split projects into categories based on whether they actively interact with people, if those people are living, if they’re collecting any new information, if the information collected can be tied to the people it comes from, if it involves people who are vulnerable in some way, what its context is, if similar research has been conducted before, etc. On the basis of the answers to all of those questions, a project will either be exempt from IRB review, qualify for an expedited IRB review, or require a full IRB review.

Under these guidelines, projects like UBiome’s (including 23andMe) are exempt. In their most basic conception, they don’t meet any of the qualifying standards (they’re not federally funded, clinical trials, or intended for peer-review). Because they’re generating a lot of data, though, 23andMe and now UBiome have decided to look into IRB approval so that they can publish any research they conduct. In an unprecedented (though long time coming) move, 23andMe consulted an independent IRB and was granted exemption from review based on the facts that their analysts do not directly interact with participants nor have any of their personal identification information. They developed a new informed consent process for their participants, but nothing really changed about their practices. UBiome doesn’t specify in the article or on their site what IRB process the went through, but it it’s likely the same as 23andMe.

So while both companies have technically met their legal requirements (or gone above them, as they point out), these cases raise a lot of questions about the IRB procedure. The research world is definitely changing, and ethical review needs to update with it. In their article, the cofounders of UBiome give a short list of what they’d like to see in the “IRB 2.0”: a relatively cheap, publicly available “mini-IRB” that could cover human research in instances that don’t require full IRB review (both 23andMe and UBiome had to pay for independent review, since they did not have their own IRB), more transparent IRB protocols (UBiome wanted to be able to see IRB proposals for similar projects in constructing their own), and public input on the development of a new IRB process.

The first concern is legitimate, but tricky. The government required IRB approval, but it doesn’t actually provide any IRBs. Many major research organizations, like universities, set up their own with staff & government approval. If an organization doesn’t want to maintain its own IRB, or doesn’t have the resources (like UBiome), they apply to an independent one. It would certainly facilitate research if the process was less expensive. The review would be just as thorough, but start-ups would be able to get a foot in the door without having to go through UBiome’s controversial solution, which was to raise funds before any review took place. The authors also lamented the amount of time they devoted to IRB review, but I don’t see the need for a change there. In the process of a review, the IRB can outright reject your proposal, which is unlikely if you took the time to fully understand the requirements and explain them in your proposal (as most researchers would), or send it back with comments for revision. If you adequately update your proposal, it can be approved. (There are obviously experiments that would not pass IRB standards, but the application process should make it pretty clear to anyone taking the time to write a proposal that their plan was unacceptable.) We can’t make individual IRBs work faster without sacrificing rigorousness, but we could make more IRBs. This could work hand-in-hand with the idea for “mini-IRBs,” which would then relieve the financial and time strain on human biology-oriented start-ups. The flip side, of course, is that more IRBs, even if “mini,” require more funding, especially if they won’t be charging as much. I’m not familiar enough with litigation over IRB proposals (I’m assuming the authors meant other researcher’s proposals, and not the working protocols of IRBs, which are available in excruciating detail) to comment on making them available to the public, but scientific research highly values transparency so this seems like a valid point. The call for public involvement in the formation of new IRB procedures, especially highlighting the most relevant groups in contemporary science, is crucial, and I think all of these issues show the need for a discussion & changes.

However, I think the concerns raised go beyond what these companies describe, and to me those are the more problematic ones. First is the fact that while the 23andMe’s or UBiome’s analysts & researchers may qualify for IRB exemption because they used anonymized data, the company as a whole has access to everything. It may be better practice to require such organizations to receive IRB approval on the basis of the company’s operations as a whole, while any individual research projects can be exempted (or not) depending on the data they use. Additionally, both 23andMe and UBiome provide (encourage, in UBiome’s case) analysis for children of essentially any age. There is a massive debate going on the the scientific community right now about genetics research in children, who are a vulnerable group because of their impaired ability to give informed consent. There are good arguments to both sides (and everywhere in between), but UBiome & 23andMe completely sidestep the ethical considerations, and end up erring on the side of abuse. UBiome asks that parents get their child’s assent (a standard practice in research involving intellectually impaired participants, where the participant gives their assent to participate and their guardian gives consent), but oddly only for children over 13, and this still raises questions about the adequacy of informed consent. In order to be ethical, assent from a vulnerable participant must be as informed as possible and uninfluenced by anyone else. Companies like 23andMe and UBiome have no way of verifying that a child assented at all, let alone in a manner up to ethical standards. This is very discomfiting, and I think strengthens the need for organization-wide IRB review. It’s not all bad, though. Many people who participated in 23andMe’s research surveys were excited to put their data to use, and interested in making that data public. IRBs have so far been concerned with keeping personal data as private as possible, so I think this situation would be an interesting new practicality for them to work out. However, I still think this should be conducted within the context of company-wide IRB review. There just needs to be new considerations for this type of data.

The other fundamental scientific units - distance, mass, luminous intensity, electric current, amount, and temperature - have been similarly redefined using more precise measures. These fundamental units are the types of unique measurements that can then be used to denote any other measurement, such as how distance (a fundamental unit) and time (a fundamental unit) can be used to express speed (a derived unit). As of today, however, temperature is still missing a definition that relates it directly to its physical source. Temperature’s current definition uses a property of water (its triple point - the temperature and pressure at which all three phases of gas, liquid, and solid exist simultaneously, which is hard to imagine from everyday life, but does happen) to anchor one fixed point of the temperature scale, the other being absolute zero, where atoms do not move at all. However, as researchers have been expanding work in very hot (like stars) and very cold (near absolute zero) temperatures, they’re noticing that this scale doesn’t quite work. While the triple point of water is a physical property, and scientists have carefully controlled the atomic properties of the water used to create the measurement, it’s simply not reliable enough to provide a really precise definition. It’s like trying to time an Olympic sprint with a seconds being 1/86400 of the time from one sunrise to the next.

To fix this problem, researchers at the National Physics Laboratory in Britain are developing a new definition of the temperature scale. They’re using a small amount of pure argon gas to more closely measure the relationship between the energy of atoms and temperature. It’s still a work in progress, but they’ve already broken new ground on the precision of the temperature definition. Their new measurement has almost halved the uncertainty or inaccuracy in the previous standard. For a little perspective, the temperature scale most people use in their daily lives is accurate to about 1 to 5 degrees; this one goes to 0.0000007 degrees.

Because organ donation is so difficult, it would be fantastic if we could get around it. That would mean that doctors could replace a malfunctioning organ with a new one made up of your own cells. Since they’re physically your own cells, they have your exact identification signals, and there’s no need to worry about the transplant being rejected. The obvious problem with this approach is that our bodies don’t just grow new organs. When they’re very young, our cells are unspecialized, and can become any type of cell we need. Those are stem cells. As our bodies develop, the stem cells specialize and lose their ability to form different kinds of cells. As a result, we can’t just take some skin cells and move them into a kidney and expect them to re-specialize and function like kidney cells.

Embryonic stem cells, the most famous ones, are found in very young fetuses and can develop into any cell in our bodies (all of our adult cells were once a limited number of embryonic stem cells). While adults do still have stem cell reserves, they’re in small amounts and already somewhat specialized, and so far scientists haven’t been able to grow them outside the body as well as they’d like. Recently, however, scientists have worked out a process that can turn certain types of our specialized, adult cells, called fibroblasts, into what are called induced pluripotent stem cells. This is amazing. We can now (rudimentarily, but we’re getting there) take cells from under the skin, put them through this process, expose them to factors that tell the new cells they should become trachea cells, and they will.

The problem is all those important things besides the cells themselves. We can grow trachea cells, but we can’t grow a whole trachea. That’s where this cutting-edge research comes in. With traditional mold-based processes or more modern 3D printing, researchers can custom build synthetic scaffolds to provide a starting point for the new organ. This has been particularly successful with tracheas and bladders - relatively simple, hollow organs. In another approach, scientists have developed a very delicate process that allows them to remove all of the cells from a donated organ without damaging the rest of the structure. Instead of printing a synthetic scaffold, this process uses a biological one, which can then be seeded with the newly grown stem cells. Researchers are trying to get this process working for complex organs like the heart, where synthetic printing doesn’t work. It’s still very difficult because part of the complexity in these organs is in their cells. Hearts have muscle cells, nerve cells, regulatory cells, and more, that all have to be in exactly the right place and function in exactly the right way in order for the heart to do its job.

It’s a difficult project, but in the end it will mean we no longer need to closely match donor organs to recipients. Instead, we’ll be able to either print our own organ or use only the non-cellular components of a donor organ. This will completely revolutionize organ transplants, and relieve the incredible stress put on the current system with waiting lists, testing potential donors, and time-sensitive organ transportation.

This is interesting and practical knowledge that will help hone our understanding of and interventions with autism, but it also exposes one of psychological testing’s deepest flaws. Psychology comes up with theories that explain behavioral phenomena in humans - young children don’t understand others’ perspectives, so we sometimes see them acting in ways that seem strange given that we do understand others’ perspectives, and we call this understanding theory of mind. Researchers then come up with simplified, controllable ways to test out these theories and learn more about them - they developed the Sally-Ann task. Researchers then use that task as a definitive test for the theory - if a child fails the Sally-Ann test, he or she does not have theory of mind.

The problem is that these tests are developed somewhat in isolation. If theory of mind was the only thing that varied in children than the Sally-Ann test would be a perfect measure of it. But, of course, there are other psychological differences, and some of them can impact what answer a child gives, independently of their theory of mind capability. In the study described in the Scientific American article, the researchers show that a number of autistic children who fail the Sally-Ann test do, in fact, have theory of mind, they just don’t apply it to the Sally-Ann test because they’re not motivated by its social incentive. Rather than giving a definitive yes-no insight on a child’s theory of mind, the Sally-Ann test actually gives the cumulative result of an interaction between theory of mind and social motivation, and probably other factors as well. Because social motivation was assumed as a given by the researchers who developed the Sally-Ann test, so no considerations were made for it. In normally developing children that might not matter, but it’s very important in autistic children. Because psychological tests are so hard to control - developers can only account for the factors they know of and consider important, and there’s no way to physically determine them - they are susceptible to problematic confounds.