Notable papers from the last year, according to some guy on the internet.
This is from Codon, a weekly newsletter about the bio + tech advances and ideas enabling a brighter future for humanity. Subscribe.
“History never looks like history when you are living through it. It always looks confusing and messy, and it always feels uncomfortable.”
— John W. Gardner
Many writers who I deeply admire have written convincing polemics on the “stagnation of progress.” Science is getting harder, discoveries are getting smaller (on average), and both of these things are bad.
If you take 93 physicists from the world’s top departments and ask them to compare a Nobel Prize-winning discovery from the 1910s to another from the 1980s, they’ll usually say that the earlier discovery was more important.
It’s also harder to make technological leaps today compared to fifty years ago. Doubling the number of transistors on a computer chip (aka Moore’s Law) requires 18 times more researchers than it did in the early 1970s. And an academic paper, published today, is less than half as likely to get cited in a U.S. patent than a paper published just 30 years ago.
These shrinking returns come in the midst of record highs.
More than 50,000 science and engineering PhDs are granted each year in the United States; that number was less than 10,000 in 1960. Federal funding for science is also (basically) at an all-time high. The total number of scientific papers is ramping up at an exponential clip, but the average number of authors on each paper has roughly quadrupled in the last hundred years.
These are all imperfect metrics of biology’s progress, for one reason or another. But there is a lot of scattered evidence to suggest that science is getting less bang for its buck, even if it feels like progress has never been faster. I’m convinced — based on anecdotal evidence, really — that biotechnology is not stagnant if benchmarked in terms of Nobel Prize-worthy discoveries per year. It’s just that a Nobel Prize is given annually for a single advance, and there’s no shortage of amazing papers waiting in the queue to win the top prize. So, of course, not every worthy person wins.
Oh, and if you take the Physics survey results that I mentioned earlier, and you do the same experiment for Medicine and Chemistry, the results are flipped — the relative importance of discoveries from the second half of the 20th century outperforms the first.
Stagnant or not, biology has massive inefficiencies. It could — should! — move much faster.
Many papers take more than a year to be published, suspended in a digital limbo under the careful eyes of journal bureaucrats. The lengthy wait probably isn’t worth it, either — peer review is often useless and a lot of bad science makes it through anyway. Lots of great ideas also never make it to a published paper because, well, they never get funded. NIH grant reviews are horribly inconsistent. If you give 25 grant proposals to 43 different reviewers, their inter-rater reliability (a measure of how consistent scores are) is basically zero, even if the grants had already received funding and high marks from previous review panels! Unfunded grants, given to the same reviewers, scored just as well as the funded ones.
And what about the biological literature? That, too, is a mess. Scientific papers are “unreliable by commission,” wrote Sam Rodriques in a recent essay, and “unreliable by omission.” In other words, some studies (a small fraction) are simply made up. Professors or students fudge numbers to get their papers published and, if anybody calls them on their bullshit, it could be years before an editor issues a retraction.
The larger problem, though, is probably sins of omission. Scientific journals like to publish positive results, so most negative findings never meet the pages of a glossy journal. If someone runs experiments that show Drug A binds to Receptor B, they’ll publish it — but neglect to mention that Drug A doesn’t bind to Receptor C through Z. Null results rarely trickle down to scientists, many of whom have had the painful experience of trying an experiment again and again for months, only to later find an obscure paper from the 1970s that made all their hard work moot.
I’m telling you these sad things — stories of stagnant science and horrid inefficiencies — to make a point: It’s amazing that biology has made so much progress in the last year. Biology’s advances are a testament to the human desire to explore, to fail, and then to continue on. This newsletter celebrates humanity’s achievements and highlights the top ten advances in biology from the past twelve months.
Everything in this list comes from a previous issue of Codon, so I hope you’ll point out my errors and omissions in the comments. My criteria for inclusion were simple: The list includes only papers that were posted to bioRxiv or published in a journal during 2022. I’m not including company achievements — which really could be its own list — unless they published the data. Some of the items on the list include multiple papers because a whole subfield moved forward really fast.
The biggest danger in putting together a “top ten” is that it’s completely subjective, there’s a lot of great stuff that didn’t make the cut, and lots of other writers have already done it (although I haven’t yet seen a list specifically for biology). Noah Smith put out his techno-optimism list for 2023, which includes some brief mentions of biotechnology, and The Atlantic published their “breakthroughs of the year” and then promptly slapped that thing behind a paywall.
There’s a need for an entire list devoted to biology, though, because a lot of great stuff has happened that was largely overshadowed by AI. The Human Genome Project (which started in 1990) was finally finished this year, as a huge team of scientists filled in the last 8% of sequence gaps. And scientists also made mirror-image DNA using a mirror-image DNA polymerase and discovered a new class of CRISPR proteins that cut proteins instead of genes. Machine learning is making big waves in protein engineering, as an algorithm was used to engineer an enzyme that can break down PET plastics faster than anything found in nature.
I’m not sure whether this year’s progress has any bearing on whether biological progress is stagnant as a whole. But one thing holds constant: Biology keeps getting weirder, and I’m here for it. ◼️
10. Toward a Synthetic Cell
Building a synthetic cell from purely chemical constituents is the Holy Grail of biology. It’s a feat that, if achieved, would prove that we understand the broad contours of how life works with enough detail to recreate it in the lab. It’d also be a starting point toward custom cells that can detect toxic pollution or manufacture medicines, without being a biosecurity risk or running risks of infection.
Several papers moved the needle on synthetic cells this year, but two spring to mind. The first, a preprint on bioRxiv, is apparently the first demonstration that ribosomes — large proteins that make other proteins — can be made outside of a living cell. It’s an important starting point for producing proteins de novo within synthetic cells.
We also made important strides in synthetic cell division. One study reported that synthetic division rings could be made in fat bubbles using just five proteins! When these proteins came together, they constricted and exerted forces on the bubbles (GIF below).
9. Better Base Editors
Back in 2016, a Nature paper reported the first base editor protein that could swap a ‘C’ for a ‘T’ in DNA, without cleaving the genome in two. Its importance was immediately clear — lots of serious genetic diseases are caused by a single base swap in DNA, and now those mutations could be fixed.
Base editors have come of age. These gene-editing proteins entered clinical trials just five years after their invention — a record pace — with ongoing trials by Verve, Beam, and other companies targeting sickle cell disease, high cholesterol, and Stargardt disease.
One of the main challenges of base editing is that it’s tricky to edit the mitochondrial genome, though, and many genetic diseases stem from there. Back in 2020, one paper showed that a base editor could successfully get into the mitochondria and make ‘C’ to ‘T’ changes, but its promise was short-lived. In May of this year, a follow-on paper said that, actually, these mitochondrial gene editors induce “substantial” off-target mutations, meaning that they edited more places than intended and probably do more harm than good.
But innovation is not stymied by failures. At least three more papers came out this year alone that have made mitochondrial base editors smaller or more accurate. It’s never been easier to get a base editor into the mitochondria or edit its DNA. I’m sure we’ll see some clinical trials specific to mitochondrial-linked diseases soon.
8. Phages Go Viral
This was a good year for phage. It’s fitting, really, because it’s also the 100-year anniversary of Felix d’Hérelle’s famous experiment. In 1922, the Parisian microbiologist showed that phages — little viruses that infect bacteria — could eradicate “dysentery and other bacilli” in rabbits and small animals.
We’re now living through a quasi-Renaissance of clinical therapies, and phage is pretty regularly being used to eliminate infections where antibiotics have failed (especially in European hospitals). A few weeks ago, a European team of scientists saved a toddler’s life using an experimental phage therapy. After an organ transplant, the young boy got a drug-resistant infection that wouldn’t clear up with antibiotics. More than two years after treatment with a custom-made phage cocktail, the child is healthy at home.
And in May, a team in Denver used phages to treat an M. abscessus infection in a boy with severe cystic fibrosis. The phages held the infection in check for more than one year until a donor lung became available. I’m sure we’ll see custom phage therapies enter the clinic in 2023.
7. Hemophilia Gene Therapies
Hemgenix, a gene therapy for hemophilia B, earned FDA approval a few months back. It will cost $3.5 million, making it the most expensive medicine ever (not good). It is safe and effective in patients for at least two years, according to phase III trial data.
Progress on a gene therapy for hemophilia A — caused by mutations in a different blood clotting protein, called VIII — has been slower. A study from May showed that a gene therapy targeted to the liver worked pretty well in non-human primates, causing a “>10-fold increase in overall factor VIII output.” And a phase I/II clinical trial with 134 participants, published in March, showed that a gene therapy delivered using an adeno-associated virus, also to the liver, caused a drop in bleeding events for people with hemophilia. It came with a lot of side effects, though; every trial participant had at least one adverse event. Still, hemophilia was one of the ‘low-hanging’ targets for gene therapies and these are useful acid tests for other targets in the pipeline.
6. Synthetic Embryos
Stem cells, harvested from mice, were used to create “embryo-like structures” this year, with working intestines, a beating heart, and the start of a brain — no sperm or eggs required.
An initial paper, by Jacob Hanna’s group in Israel, was published in Cell in August. Those authors have since formed a company, called Renewal Bio, “that aims to grow human synthetic embryos to provide tissues and cells for medical conditions,” according to coverage in The Guardian. A preprint by a team from Cambridge University and Caltech was also released on August 2nd, and showed that these “synthetic embryos” accurately recapitulate “developmental events from embryonic day 5.5 to 8.5, including gastrulation, and formation of the anterior-posterior axis, brain, a beating heart structure, and the development of extraembryonic tissues, including yolk sac and chorion.”
Synthetic embryos — which only form correctly about 0.5 percent of the time — are potentially useful for a few reasons. These structures could be used to study how organs form during development, for instance, but also to test medicines without using real embryos.
5. Cellular Reprogramming
A massive highlight from this year: The first demonstration that chemicals alone can reprogram human cells back into stem cells. The breakthrough took more than a decade of work and requires eleven different chemicals and between one and two months of work, so it will need some fine-tuning. The same group did this in mouse cells way back in 2013, but the process for human cells is much more difficult.
This also isn’t the first study to reprogram human cells; that honor belongs to Shinya Yamanaka who, in 2006, reprogrammed iPS cells by expressing four proteins (the now-famous “Yamanaka factors.”) In the decades that followed, other groups used viruses or mRNA to reprogram cells. But this chemical-only approach stands out because it’s simple to use in vitro and the chemical cocktails can be delivered into the body using IVs — no gene-editing required. The approach provides a relatively simple tool to generate human pluripotent stem cells that can be used in regenerative medicine.
4. Plants Got Easier to Engineer
We are living through the Anthropocene, a terrifying moment in human history in which man-made materials weigh more than all biomass on Earth and continue to double in weight every year.
Of our existing biomass, an estimated 83 percent is held by plants. If we want to scrape our way out of this planetary dumpster fire, then, we’ll probably have to engineer plants: Make them capture more carbon, supercharge their photosynthesis, grow more food, something!
Fortunately, it’s never been easier to do that. For decades, synthetic biologists have engineered bacteria and mammalian cells with increasingly complex genetic circuits, even as plants went largely neglected. This year, two important advances shifted the balance.
The first is that gene-editing technologies, like CRISPR, have now been optimized to work better in plants. Prime editors, for example, are ‘search-and-replace’ gene editors that can insert, delete, or swap DNA. A new Plant Prime Editor is up to 3.4-fold more efficient in plant cells compared to other prime editors, and was used to swiftly make rice plants tolerant to herbicides in the laboratory.
But the bigger advance was this: A Stanford team released an entire genetic toolkit to “program” plants in much the same way that we program bacteria. The new toolkit includes a host of synthetic promoters and transcription factors that can be used to control gene expression in plants. These genetic parts were used to construct gene circuits capable of Boolean logic operations in N. benthamiana and Arabidopsis. The authors also built logic gates that could control gene expression levels in plant roots to control their lateral densities.
3. Going Carbon-Negative
A notoriously difficult part of biology is that it’s hard to scale. It’s one thing to engineer a cell that can turn sugar into a cancer drug in a test tube, but an entirely different beast to do the same in a thousand-liter bioreactor.
That’s why companies like LanzaTech are exciting. They’ve actually scaled up biology at industrial factories. They have pilot facilities that are recycling waste carbon from factories into fuels and chemicals. These plants could make hundreds of billions of gallons of fuel each year and are already scrubbing emissions at levels equivalent to taking thousands of cars off the road each year.
A few months back, scientists from LanzaTech and Northwestern University reached another milestone: Using an engineered autotroph, called Clostridium autoethanogenum, they produced acetone and isopropanol at industrial scales in a process that had a negative carbon footprint. “Unlike traditional production processes, which result in release of greenhouse gases, our process fixes carbon,” they wrote in the published paper. It’s a huge step forward for biology at scale. The engineered microbes can ameliorate pollutants and produce chemicals without dragging down the atmosphere.
2. Xenotransplantation Made Reality
This will go down in the annals of human history as the “Year of Xenotransplantation.” With a history stretching back to the mid-1960s — when the French surgeon, Renè Kuss, transplanted a pig kidney into a person only to have it swiftly rejected by the patient’s body — xenotransplantation has been a field of failure. The most famous example is probably the 1980s surgery of Stephanie Beauclair, or Baby Fae, who received a heart from a baboon and died shortly after.
We’ve gotten a lot better at taking hearts from pigs and putting them into humans over the last forty years, though, because gene editing techniques are improving. It’s easier than ever to find all the nasty proteins in a pig’s cells which cause our body to reject the organ, and systematically eliminate those protein-coding genes from the genome. But progress is never without pitfalls.
On January 7th of this year, a team of surgeons at the University of Maryland transplanted a pig’s heart into 57 year-old David Bennett Sr. He died two months later.
Dejection was followed by optimism when, in May, two pig kidneys were transplanted into brain-dead people and monitored over a 54-hour period. Both kidneys produced urine and biopsies did not show any signs of organ rejection.
The heart transplant likely failed because the organ was infected with a pig virus, but the authors of the kidney transplant study detected no such viruses. Pig kidneys may soon make their way into the living.
1. We’ve Never Been Closer to Ending Malaria
“Malaria has killed half of all the people who ever lived,” if you believe this 2002 Nature article by John Whitfield (the claim was uncited.) Even if the claim is false, mosquitoes have collectively killed billions of people throughout human history. The insect bastards took out Alexander the Great, Alaric the Goth, and Dante. And still, in 2021, malaria killed about 620,000 people. It’s shocking, sad, and preventable. But this year was one of amazing progress.
One malaria vaccine, called R21/Matrix-M, was tested in four African nations in a phase III trial. Its overall efficacy was 75 percent in young children. That’s fantastic news. A phase II trial in Mali also tested an antibody, called CIS43LS, against P. falciparum infections in healthy adults. A single shot had an efficacy of 88.2 percent over a six-month span.
The only immunogens that are more than 90 percent effective at preventing malaria, though, are sporozoites, a spore-like part that forms during the mosquito’s lifecycle. These sporozoites are injected into the arm and used as a vaccine, but it’s obviously difficult to collect these things from actual mosquitoes.
Earlier this month, researchers created “hundreds of millions” of sporozoites in the lab; no mosquitoes required. They basically recreated the entire “P. falciparum life cycle from infectious gametocyte to infectious gametocyte without mosquitoes.” The new technique will help create vaccines against malaria much faster and more cheaply.