Reflections on the Song D614G, an Evolutionary Tale
Bette Korber, April 9, 2023.
This is a much-too-long version of the short essay that will accompany my
song on YouTube when it is released later this week.
D614G, an Evolutionary Tale is just a personal expression of my feelings around the events that took place in 2020, shared through a song. (Yes, scientists do have feelings – who knew?). It is also an attempt to share this bit of science and how the science was actually done with both scientist and non-scientist friends and family. It does not reflect the views of my employer or even my co-authors on the paper, just my own perspective.
This song is the story of this story (1):
Tracking Changes in SARS-CoV-2 Spike: Evidence that D614G Increases Infectivity of the COVID-19 Virus. Korber et al. Cell 182(4):812-827 2020
This paper presented evidence that the infectivity of SARS-CoV-2, the virus that causes COVID-19, was increasing due to a mutation called “D614G”. It also showed that a variant with a fitness advantage like D614G could sweep through the entire planet in a matter of a few months.
The study was released as a preprint (2) in the spring of 2020 at a time when the prevailing scientific view was that SARS-CoV-2 was not yet evolving under selective pressure (3). Since our work indicated otherwise, it was the subject of a heated controversy. Although many scientists agreed with our interpretation of the data, some disagreed and did so with reasoned discussion, and there were others who bordered on hostile, and in their zeal to express their view were sometimes willing to weigh in with some nonsense.
When you are working on the edge of the known and are trying to incorporate a new idea into a framework of scientific thinking, of course there will be disagreement. Before novel work is widely accepted, a series of confirmations to check whether the work can be repeated, as well as new ways of testing a hypothesis, need to be advanced. That is as it should be. But in the era of social media, with reporters amplifying casual arguments and speculations from tweets and blogs, scientific disagreement has not only become very public, but also sometimes charged and acrimonious. This amplification of discord neither serves science nor elevates the public’s opinion of scientists; we have seen it happen with SARS-CoV-2, with both the D614G story and then again, quite recently, in the continuing debate over the origins of the virus in humans in Wuhan.
As SARS-CoV-2 was beginning to spread globally, many scientists early on voiced the view in social media and in news stories that the virus was not evolving under positive selection, and that the mutations observed were not biologically important (3,4). I suspect people may have held this view at least in part because they wanted to shore up hope, both to comfort themselves and a frightened public, with the reassuring thought that things would not get worse. Of course, things ultimately did get worse, as major case waves were to follow the emergence of new variants such as Delta and Omicron. Early on there was also a hope that the vaccines that were urgently under development throughout 2020 would continue to provide effective immune responses against viral infections for years to come.
What is clear to everyone now is that SARS-CoV-2 has evolved, and continues to evolve, both to be more infectious and to escape neutralizing antibody responses (a special kind of antibodies that are very good at blocking new infections). The variant du jour (as of this moment in April 2023) is an Omicron recombinant sublineage called XBB.1.5, and it has both of those selective advantages relative to what came before it (5,6). Still, COVID-19 vaccines and natural infections continue to elicit valuable immune responses that help us on many fronts. Vaccination can enable better control of the virus if you are infected, which helps protect from severe disease and death, and it also reduces the chance of infecting others. Vaccines can be more potent when more contemporary variants are included, so we will need to continue to adapt our vaccine strategies to accommodate viral adaptations as they unfold. Now, after more than 3 years of living in a world with COVID-19, we the people have built up useful immunity to SARS-CoV-2, and we have immune protection based on both prior infections and vaccines to help us going forward in this “post-pandemic” world. But the virus continues to evolve within us as a species, and so will we need to keep an eye on it in the future.
But back in early 2020, some of the arguments that the virus wasn’t evolving under selection (7) were neither evidence-based nor even logical given what we understand about pathogen introduction into a new host (that “new host” would be us, human beings). Examples of such arguments were, “there’s no evolutionary pressure on the virus to transmit better” and, “it had already found its best way of being a [human] virus.” It is hard for me to fathom why people would have assumed these statements to be the case a priori. Another odd statement that was often repeated when people were trying to refute the D614G relevance was, “a single mutation could not possibly cause a phenotypic change.” This is simply incorrect. There are vast numbers of examples of exactly this in the scientific literature, and D614G is just one case of very, very many. A well-known example among those in the HIV-1 scientific community is that a single amino acid change in HIV-1 Env can dictate whether a particular virus replicates well in humans or in monkeys by modulating how well the virus can interact with the human or macaque form of its receptor, CD4 (8) — a story with some parallels to D614G in Spike and its human receptor ACE2.
Others used semantic arguments to discredit our work, specifically targeting the D614G story with their criticism and effectively sidetracking the D614G discussion away from the science and instead to a muddled story about phrasing. One example of this was the buzz that we shouldn’t use the word “strain” to describe a new variant, as we did in our bioRxiv preprint (2). Strain, in just the way we used it, is common usage the HIV-1 field; this word was already being used by many others besides us in the newly emerging SARS-CoV-2 field, which was why we adopted it, but we became singled out as the focus of objections to its use. Some held that the term strain should only be used after a scientific review board had declared that a new strain classification was merited. That in itself is a legitimate difference of opinion. But there were no such review boards in 2020, and this was not a widely held convention, and many other scientists were using "strain" just as we did, as essentially synonymous with the word "variant". (A quick search of PubMed (2023/04/08) reveals that “strain” and “SARS-CoV-2” come up together in 5,283 scientific papers). We ultimately switched to the word variant not because we felt strain was inappropriate, but to avoid having to engage in this distracting discussion. Happily, this particular discussion has, for the most part, been relegated to the past, and people most often use the term variant now.
Once our paper was formally published in a peer-reviewed journal (1), the Los Alamos National Laboratory invited us to respond to these and other spurious arguments, and we did in our FAQ (9).
It is important to note that we did not publish our first version of the paper in the bioRxiv preprint server until we had experimental support of our bioinformatics/statistical findings. We waited for experimental support. This first support was the clinical data from Dr. Thushan de Silva that showed there was a statistically significant increase in viral RNA in the upper airways from people who had D614G viral infections compared to those who had ancestral viral infections (2). Nonetheless, some people chose to ignore this evidence or to brush it off. Meanwhile, experimentalist colleagues who worked with us on our study, Dr. David Montefiori and Dr. Erica Ullmann Saphire, continued to acquire evidence confirming the importance of D614G, so that by the time we published our Cell paper, they had also proven experimentally in vitro that D614G was more infectious (1). Other laboratories also confirmed both Thushan’s clinical results (10,11) and David and Erica’s in vitro results (10,12) and these independent confirmations were also available by the time we published in Cell. Still there were those who said that our observations were not supported and/or were not relevant.
Additional confirmations came soon after our paper was released. We and others published structural studies showing mechanistically how D614G altered Spike and made it more infectious (13-16) including some particularly beautiful and informative protein structures that came from Dr. Priyamvada Acharya's group at Duke University (16) and proteins dynamics modeling from my colleague Dr. Gnana Gnanakaran at Los Alamos (14). By the autumn of 2020, Dr. Ralph Baric had shown that the virus was more infectious in animal models, a key bit of data many were waiting for (17). Despite the growing evidence demonstrating the importance of the D614G (bioinformatic (1-2), clinical (1,2,10,11), in vitro (1,9,13), structural (13-16), and in vivo in animal models (17)), still the controversy was thriving in September of 2020 (18). (If you’re interested in details see section at the end.*)
There were a number of reasons I felt our results were important to document and share. Simply understanding the evolutionary potential of this virus and discovering the remarkable speed with which it could transition globally was important. This understanding would have allowed realistic expectations and better knowledge of what to expect going forward. It could have helped in making the public health case for social distancing and masking in a time when we still had a year to go before the general population could get vaccinated. But of course, we will never know for sure what the impact of greater clarity early on during the pandemic would have been.
Deepest thanks
I would like to thank some people, as well as feathered and furred beings, who helped with this work and with the communication of this work, or who kept our minds and spirits going in times of despair in COVID’S shadow.
I. Dark Birds:
If you are wondering how the dark birds came to roost in the song, throughout 2020 crows and magpies would swoop in for breakfast or an afternoon snack on our porch, winging with utter grace through the lonely space of COVID-19 lockdown. There were 3 crows looking after a nest in the canyon where we live. In the late afternoon they would do performance art on the winds stirred up by the canyon. In New Mexico, ravens are always about. They all grew to feel a bit like extended family, so I decided to indulge myself by having a dark bird symbolize my husband James and myself in the song.
Also, this is a song about evolution, “the song that is never-ending,” so the essence of the idea of Darwin’s finches flutters into the 3rd verse and messes around with the viral RNA.
II. GISAID::
The model for viral sequence data-sharing
that GISAID offered was new, and it made variant discovery in real time possible
(22-24). GISAID continues to enable my small group at Los Alamos, and so many
others, to track this virus. I do not think this has been appreciated enough by those in my field who use this data constantly in
their work. But GISAID has been controversial from the earliest days in 2020, and that controversy remains very heated though this moment in 2023. Still it is what enabled tracking SARS-CoV-2 evolution globally in close to real time throughout the history of the pandemic. My hope is that somehow, despite the history and acrimony, we in my field will be able to find a better way forward together, that we can learn and, well, adapt. SARS-CoV-2 continues to evolve, and future pandemics may be brewing.
The full GISAID database of over 15 million sequences was used to make several of the slides in the video.
I wrote what follows in my credits in the video, and it comes from my heart and my mind. I think of GISIAD as a community, all who contribute data and all who use it, as well as the people who enabled data sharing at GISAID.
“This 15 million sequence miracle of global cooperation enabled a dynamic response to this ever-changing virus.
Deepest thanks to:
- The people at the frontline clinics
- The sequencing labs and bioinformatics folks who processed the data
- Those at GISAID who developed the conceptual and computational framework to share the data
- Those who extract the meaning from the vast river of data that holds the story of this virus”
III. Experimental colleagues.
I’m ever grateful to my steadfast friend and wonderful colleague Dr. David Montefiori. We have worked together on HIV-1 for decades, trying to decipher the implications of the evolution of neutralizing antibody escape. We worked together on COVID-19 through the Dark Ages (the Spring of 2020); as he was working on this he was also deeply embedded in the work of developing consistent assays for monitoring vaccine response antibodies and getting FDA approval. I’m delighted and still amazed that he worked with me yet again to create this music video; he came out from Duke to New Mexico for the recording, added some chimes to touch the music directly, and he provided pictures of his lab where the actual experimental work was done. David also wrote his own notes about our time in 2020 to co-release with this song (25).
I’m also deeply grateful to Dr. Erica Ollmann Saphire; when she confirmed our findings regarding D614G experimentally she didn’t try to scoop us or compete with us. Instead, she very graciously offered to work with us, and shared her data; her lab was a few days ahead of David’s, but they were very close in time. They each had demonstrated independently that D614G Spike was more infectious in vitro. Erica also offered David and me invaluable help in terms of her insights and perspective as we were writing our Cell paper.
Thanks also to Dr. Thushan de Silva, who responded to our request to work together. When I shared my analysis with him and he saw how D614G was on the rise, he figured out how to test our hypothesis using his samples. He did this at a time when many others were saying we were wrong, and so I think it took some spine on his part to stand with us. I loved working with him and his wonderful team – they also kept us thinking straight with regard to the recombination story we were working on at the time. That analysis was in the preprint (2), and Thushan and his team improved it; but our reviewers felt the recombination story, while well substantiated, was too complex and distracted from the main point, which was D614G, so they recommended that we not include it in the Cell article (1).
Also, I remember with gratitude some scientists who were working on the problem at the same time we were, thinking as we were, and who were open and gracious in sharing their data as they were reporting it. Two fine people come to mind in particular: Dr. Michael Farzan (12) and Dr. Jeremy Luban (13).
III. Science Communicators.
America and in fact the world owes both Dr. Anthony Fauci and also writer and commentator Laurie Garrett gratitude and respect for their science communication throughout the hardest times of the pandemic. They were fearless, reasoned, spoke with clarity, and they did their best to keep truth forward. I know both of them considered this D614G work carefully, and they understood the implications of D614G the first time they saw it laid out. Laurie sent me an email on May 23, 2022, saying “I heard/watched your HVP Zoom talk. No doubt about it, the D to G switch is a big deal.” I sent our paper to Tony as soon as it was accepted. I discovered from my friend Joanne Topol, who had recorded the broadcasts and sent them to me, that Laurie and Tony both had shared this science with the public in an accurate and calm way through their television news updates as soon as it came out in the peer-reviewed literature.
The journalist Ralph Vartabedian broke the story in the Los Angeles Times when our preprint first came out (26). He read the paper himself and had his head around all of the details (not a trivial point, as I’ve discovered since that not all reporters attempt this when covering new science studies), and he got the science right. Joel Achenbach of the Washington Post wrote a great article that retrospectively set the record straight in Oct of 2021 (27). I so deeply appreciated his remembering the controversy and making the effort to do this.
IV. Friendly Skeptics.
I’m also grateful to those who were initially skeptical and reserved judgment on the impact of D614G until they could test for the increased infectiousness of D614G in their own labs and in their own way, but who engaged in respectful reasoned dialog with us. Dr. Ralph Baric and Dr. David O’Connor both come to mind as good examples. They were superb colleagues. David eventually found the same kinds of transitions to D614G we were observing repeated in his own data from Wisconsin, and Ralph ultimately proved D614G was more infectious in animal models in his laboratory (17). Respect.
V. Life Bringers.
With thanks to those who brought joy into our locked-down, work-obsessed world in 2020 and forced us to look up from our computers. Particular thanks to my husband, Dr. James Theiler, a mathematician/physicist who lent his programming skills and scientific insight to our SARS-CoV-2 analyses throughout the pandemic. He had retired in March of 2020, and essentially all of the work he did was as a volunteer – it was often many days a week. My friends Hyejin Yoon and Will Fischer worked incredibly hard on this project. My friend Kshitij Wagh kept our HIV-1 work going strong and shouldered that responsibility enabling me to focus more on SARS-CoV-2. Our son Sky Korber lived with us and did all of the cooking and shopping through the first year and a half of the pandemic so we could focus on work. Not only was Sky a brilliant cook, but he also kept us up on the news.
We were blessed to be able to work from home, and we only saw a few people outside of our immediate family, in particular our friends Don Medina and Val Archuletta who were just starting their own small local farm and were also starting their own beautiful family. (I have a song for Val and the farm, The Fox.) In 2020 they brought us eggs and vegetables they had grown, and came every week for a visit on the porch with their beautiful little daughter Elena, who stole my heart. Then came Storrie and Auggie. (I have a song for the kids, Three Peas in a Pod.) In 2020 they also brought us Zephyr, a little sheepdog pup, and he was just what we needed. (Surprise! He has a song too. How Zephyr McLeod Got His Name). Music is my joy and my hobby, and my musical mentor Lisa Carman zoomed in with lessons throughout the pandemic. When a song was ready I could take it up to Taos to Peter Oviatt’s Moonflower Sounds studio, shake off my lockdown blues and remember the World and the Rio Grande were still out there flowing, and I would record with him. In 2020, this meant him using an improvised outdoor studio with quilts hung on lines to dampen the sounds and to shield the mics from the wind, and the birds all around us would join in on the recording.
Our very small circle included our dear long-time friend Garret Kenyon. When the weather was fine on Sunday afternoons, we would share a dinner on the porch and watch a play or musical (instead of a regular movie, to make it special), our “Theater on the Porch”. Gar and James were the first ones to see my initial D614G work, just as James began to automate my scrappy code to enable automated global surveillance from GISAID data. We can’t remember which of the two had said it, but either Gar or James pointed out there was going to be a lot of pushback and brought up a meme, “First She discovers it, and He says, ‘She’s wrong.’ Then She proves it even to Him, and He says, ‘Well, it’s true but it doesn’t matter.’ Then when He notices people are interested, He says, ‘I discovered it, who the hell is She?’ “ When that prediction actually came to pass, it inspired the chorus of the song. As that prediction was coming to pass in 2020, my sister Dorothy Korber, my best friend Joanne Topol, and my far-away son Max Theiler would let me rant, make me laugh, and talk me down, they were phone links to sanity.
My songs are mostly just shared by family and good friends, so I think Alpha Man is quite unlikely to stumble across this song. However, if “He” does (Alpha Man is actually a composite of a few men) I suspect “He” might still say this work was either not important or attribute the findings not to our team but to someone else. But I think the citations can speak for themselves. Our bioRxiv preprint (2) as of April 10th, 2023, has the third highest “attention score” in the bioRxiv (#3 of 189,662). Google scholar noted that the Cell paper (1) was cited 3,691 times, which is quite a lot for a scientific paper. As for the importance of the D614G mutation to the SARS-CoV-2 field, a joint search on “D614G” and “COVID-19” in PubMed finds 883 scientific publications, while a more inclusive search offered by Google Scholar identifies 1,980. Alpha Man might have felt D614G was unimportant; others disagreed.
Mind you that a lot of the attention was not exactly positive attention. I was trying hard to not read the tweets or the tweet-echoes in the news, but enough of the jabber got through it was hard to live with, even though I was confident of the data and of the analysis and conclusions. But what was far harder for me during this time was that I felt the science mattered to public health in the moment, and that we needed an unflinching view of what we were up against -- not false nor condescending comfort; D614G gave us a preview of what was to come, and people were not attending to it. So, I kept speaking to the point, and it wasn’t easy. My work weeks were generally 70+ hours.
Alpha Man might suggest that this song is a misrepresentation of the
history around this story. But it is my truth (well not literal truth, James
and I aren’t actually dark birds, the lyrics are written with a bit
of poetic license). It was cathartic for me to sing my song (art heals) with my
wonderful musician friends (Peter
Oviatt, Mohit Dubey, Lisa Carman, Karina Wilson, Justin Bransford, Brent Berry, and Scott Shane) who lifted it up and transformed it
into music. Also, it was a gift to create this music video with the help of
many friends who shared their amazing images to enable me to tell this story
and share some of the science and an impression of how the science is actually
done, from both David’s and my perspective. I am especially grateful to my dear
friend Andy Black for his gorgeous animations of birds and his remarkably
accurate SARS-CoV-2 RNA double helix. (Andy and I did a cover song music video
together in 2020, Boots
of Spanish Leather, where the whole video is his animation.) Joanne Topol, Irene Owsley, and Kat Livingood were so kind in sharing their beautiful images to help me tell this story by borrowing from their vision.
Finally, while I had my own trials and tribulations through 2020, I know they were nothing compared to those who suffered directly from the pandemic. So many had to face serious illness, and some friends faced the day of their own death when their breath would no longer come. So many lost family. Those who worked on the front lines in the hospitals and clinics were epic heroes. Those who lost their livelihoods and their homes in those crazy times, or had to get their little ones through home schooling and the loneliness, all dealt with very serious hardships my immediate family was spared.
My friends who went through those sorrows, my heart was with you then, and I remember, and I love you.
-- bette
*Some further thoughts on the controversy that lingered on:
In his Nature commentary in September 2020 (18), 2.5 months after our Cell paper had come out, Dr. Ewen Callaway pointed that in our bioRxiv preprint we had an exception to the pattern that as soon as D614G was established in a geographic region or county it would start to take over and replace the ancestral form – the exception was Iceland. He omitted that we had resolved exactly why this happened in the Cell paper. We had identified 50 geographic regions where D614G and the ancestral virus were co-circulating. We found D614G was significantly increasing in 48 of the 50; there were only two exceptions, and we understood both of them. One was Iceland, and we had learned that first samples collected in Iceland were all from people getting off planes from Europe, where D614G was already common. When the team from Iceland later started sampling virus from people living in Reykjavík, Iceland’s capital, instead of people flying in, they found that the ancestral form was established there and common locally, so the sampling initially shifted backwards from D614G (in European travelers) to ancestral (in the local population). Soon afterwards, D614G replaced the ancestral form in the Icelandic population. The other exception was Santa Clara County, in California. D614G had been introduced in Palo Alto early on, and wasn’t established in San Jose until several months later. This meant that early on the frequency of D614 in Santa Clara county overall depended on whether Palo Alto or San Jose was being more heavily sampled in a given time frame. The first samples were dominated by samples from Palo Alto, later samples from Santa Clara. Once D614G had a foothold in Santa Clara, it expanded there too.
Callaway also highlighted our use of the word “alarming”, and noted that we had used it in our preprint but it was “scrubbed” from our Cell paper. One of our reviewers had suggested we should take it out, so we did. It had become a topic of debate much like the word “strain”. After the fuss over the word, some of my co-authors were indeed relieved to remove it, and I also did not want a word-usage debate to distract from the science. But in point fact, I did find it alarming that a virus with a more infectious phenotype could become globally dominant in just a few months, and I thought people should know that. I felt then and still feel today that the word was appropriate. During this period, I was informed by several Alpha Men that a Proper Scientist should never use such an inflammatory word as “alarming” to describe a scientific finding. So I found it gratifying when Dr. David Ho, a brilliant scientist that the scientific community holds in the highest esteem and who is particularly tempered in his style, later chose to use the word “Alarming” in the title of a scientific paper describing immune escape in SARS-CoV-2 variants (19). I strongly suspect David was not later the recipient of multiple condescending communications instructing him, for his own good, that he was not being properly circumspect.
Sometimes science is, indeed, alarming.
I think from the blogs and from discussion in 2020 that part of the controversy was initially fueled by positive selection not being detected in SARS-CoV-2 using tests for recurrent mutations in phylogenetic trees. It is hard to publish negative results, but still some examples got published (20). These are generally sound methods, I use them too depending on the application, but the approach was inadequately sensitive given the limited diversity of the virus in 2020 (21). Absence of proof (by inappropriate methods) is not proof of absence. D614G was not re-introduced by mutation repeatedly in many lineages; instead, the viral lineage that carried it outgrew all other lineages that were present and co-circulating at the time, in every geographic region soon after it was introduced. An analogy: we found the bloody knife at the crime scene, but our critics complained that there was no smoking gun.
1 Tracking Changes in SARS-CoV-2 Spike: Evidence that D614G Increases Infectivity of the COVID-19 Virus Korber et al. Cell 182(4):812-827 2020
2 Korber et al. bioRxiv release May 5, 2020. doi: 10.1101/2020.04.29.069054
3 Grubaugh, Nature Microbiology, 5(4), 529–530.
4 The Problem With Stories About Dangerous Coronavirus Mutations. By Ed Yong, The Atlantic, May 6, 2020.
5 Uriu K et al. Lancet Infect Dis. 2023 Mar;23(3):280-281.
6 Lasardo et al. bioRxiv. 2023 Jan 23 doi: 10.1101/2023.01.22.525079.
7 Why the Coronavirus Has Been So Successful by Ed Yong, The Atlantic. Mar 20, 2020
8 Li et al. Proc Natl Acad Sci U.S.A. 2016 Jun 14;113(24):E3413-22.
9 FAQ: Tracking SARS-CoV-2Spike mutations
10 Lorenzo-Redondo R. medRxiv. 2020; (2020.05.19.20107144)
12 Zhang et al. bioRxiv doi: 10.1101/2020.06.12.148726
13 Yurkovetskiy et al. Cell. 2020 183(3):739-751
14 Mansbach et al. Sci Adv. 2021 16:7(16)
15 Weissman et al. Cell Host Microbe. 2021 29:1(23-31)
16 Gobeil et al. Cell Rep. Movie by Bette Korber 2021 34(2):108630
17 Hou et al. Science. 2020 Nov. 12; 370(6523): 1464–1468
18 Callaway Nature 585:174 Sept 10, 2020.
21 Wang et al. Cell 186: 279–286 2023
20 Van Drop et al. Nat Commun. 2020 Nov 25;11(1):5986.
22 Khare, S., et al (2021) China CDC Weekly, 3(49): 1049-1051.
23 Elbe and Buckland-Merret. (2017) Global Challenges, 1:33-46.
24 Shu, Y. and McCauley, J. (2017) EuroSurveillance, 22(13).
25 Reflections on our COVID-19 study in early 2020, by David Montefiori,
26 Scientists say a now-dominant strain of the coronavirus could be more contagious than original. By Ralph Vartabedian Los Angeles Times, May 5, 2020https://www.latimes.com/california/story
27 The coronavirus is stillmutating. But will that matter? ‘We need to keep the respect for this virus.’ By Joel Achenbach, Ben Guarino, and Aaron Steckelberg, Washington Post, October 18, 2021.
