Found: Genes that sway the course of the coronavirus

A study of some of the sickest COVID-19 patients, such as those placed on ventilators, has identified gene variants that put people at greater risk of severe disease.

Fabio Teixeira/Anadolu Agency/Getty Images

Sciences COVID-19 reporting is supported by the Pulitzer Center and the Heising-Simons Foundation.

It’s one of the pandemic’s puzzles: Most people infected by SARS-CoV-2 never feel sick, whereas others develop serious symptoms or even end up in an intensive care unit clinging to life. Age and preexisting conditions, such as obesity, account for much of the disparity. But geneticists have raced to see whether a person’s DNA also explains why some get hit hard by the coronavirus, and they have uncovered tantalizing leads.

Now, a U.K. group studying more than 2200 COVID-19 patients has pinned down common gene variants that are linked to the most severe cases of the disease, and that point to existing drugs that could be repurposed to help. “It’s really exciting. Each one provides a potential target” for treatment, says genetic epidemiologist Priya Duggal of Johns Hopkins University. 


Kenneth Baillie of the University of Edinburgh, an intensive care physician and geneticist, led the new study, which he discussed on 2 October at an online meeting of a data-pooling effort called the COVID-19 Host Genetics Initiative. He’s hoping the results, also posted as a preprint on medRxiv, will speed treatments, although he cautions that any clinical trial inspired by the findings should wait for the study’s acceptance in a peer-reviewed journal. “Because the epidemic is progressing at such an alarming rate, even a few months of time saved will save lots of lives,” Baillie says.

In a standard approach to finding genes that influence a condition, geneticists scan the DNA of large numbers of people for millions of marker sequences, looking for associations between specific markers and cases of the disease. In June, one such genomewide association study in The New England Journal of Medicine (NEJM) found two “hits” linked to respiratory failure in 1600 Italian and Spanish COVID-19 patients: a marker within the ABO gene, which determines a person’s blood type, and a stretch of chromosome 3 that holds a half-dozen genes. Those two links have also emerged in other groups’ data, including some from the DNA testing company 23andMe.

The new study confirmed the chromosome 3 region’s involvement. And because 74% of its patients were so sick that they needed invasive ventilation, it had the statistical strength to reveal other markers, elsewhere in the genome, linked to severe COVID-19. One find is a gene called IFNAR2 that codes for a cell receptor for interferon, a powerful molecular messenger that rallies the immune defenses when a virus invades a cell. A variant of IFNAR2 found in one in four Europeans raised the risk of severe COVID-19 by 30%. Baillie says the IFNAR2 hit is “entirely complementary” to a finding reported in Science last month: very rare mutations that disable IFNAR2 and seven other interferon genes may explain about 4% of severe
COVID-19 cases
. Both studies raise hopes for ongoing trials of interferons as a COVID-19 treatment.

A more surprising hit from the U.K. study points to OAS genes, which code for proteins that activate an enzyme that breaks down viral RNA. A change in one of those genes might impair this activation, allowing the virus to flourish. The U.K. data suggest there is a variant as common and influential on COVID-19 as the interferon genetic risk factor.

Other genes identified by Baillie’s team could ramp up the inflammatory responses to lung damage triggered by SARS-CoV-2, reactions that can be lethal to some patients. One, DPP9, codes for an enzyme known to be involved in lung disease; another, TYK2, encodes a signaling protein involved in inflammation. Drugs that target those two genes’ proteins are already in use—inhibitors of DPP9’s enzyme for diabetes and baricitinib, which blocks TYK2’s product, for arthritis. Baricitinib is in early clinical testing for COVID-19, and the new data could push it up the priority list, Baillie says.

The chromosome 3 region still stands out as the most powerful genetic actor: A single copy of the disease-associated variant more than doubles an infected person’s odds of developing severe COVID-19. Evolutionary biologists reported last month in Nature that this suspicious region actually came from Neanderthals, through interbreeding with our species tens of thousands of years ago. It is now found in about 16% of Europeans and 50% of South Asians.

But the specific chromosome 3 gene or genes at play remain elusive. By analyzing gene activity data from normal lung tissue of people with and without the variant, the U.K. team homed in on CCR2, a gene that encodes a receptor for cytokine proteins that play a role in inflammation. But other data discussed at last week’s meeting point to SLC6Z20, which codes for a protein that interacts with the main cell receptor used by SARS-CoV-2 to enter cells. “I don’t think anyone at this point has a clear understanding of what are the underlying genes” for the chromosome 3 link, says Andrea Ganna of the University of Helsinki, who co-leads the COVID-19 Host Genetics Initiative.

The U.K. genetics study did not confirm that the ABO variants affect the odds of severe disease. Some studies looking directly at blood type, not genetic markers, have reported that type O blood protects against COVID-19, whereas A blood makes a person more vulnerable. It may be that blood type influences whether a person gets infected, but not how sick they get, says Stanford University geneticist Manuel Rivas. In any case, O blood offers at best modest protection. “There are a lot of people with O blood that have died of the disease. It doesn’t really help you,” says geneticist Andre Franke of the Christian-Albrecht University of Kiel, a coleader of the NEJM study.

Researchers expect to pin down more COVID-19 risk genes—already, after folding in the U.K. data plumbed by Baillie’s team, the COVID-19 Host Genetics Initiative has found another hit, a gene called FOXP4 implicated in lung cancer. And in a new medRxiv preprint posted last week, the company reports that a gene previously connected to  the effects of the flu may also boost COVID-19 susceptibility only in men, who are more likely to die of the disease than women.

Geneticists have had little luck so far identifying gene variants that explain why COVID-19 has hit Black people in the United States and United Kingdom particularly hard. The chromosome 3 variant is absent in most people of African ancestry. Researchers suspect that socioeconomic factors and preexisting conditions may better explain the increased risks. But several projects, including Baillie’s, are recruiting more people of non-European backgrounds to bolster their power to find COVID-19 gene links. And in an abstract for an online talk later this month at the American Society of Human Genetics annual meeting, the company Regeneron reports it has found a genome region that may raise the risk of severe disease mainly in people of African ancestry.

Even as more genetic risk factors are identified, their overall effect on infected people will be modest compared with other COVID-19 factors, Duggal says. But studies like the U.K. team’s could help reveal the underlying biology of the disease and inspire better treatments. “I don’t think genetics will lead us out of this. I think genetics may give us new opportunities,” Duggal says.

How smart devices are exploited for domestic abuse

Another explained how her partner used Amazon’s virtual assistant to monitor her via a function that lets users remotely connect to enabled smart speakers .css-yidnqd-InlineLink:link{color:#3F3F42;}.css-yidnqd-InlineLink:visited{color:#696969;}.css-yidnqd-InlineLink:link,.css-yidnqd-InlineLink:visited{font-weight:bolder;border-bottom:1px solid #BABABA;-webkit-text-decoration:none;text-decoration:none;}.css-yidnqd-InlineLink:link:hover,.css-yidnqd-InlineLink:visited:hover,.css-yidnqd-InlineLink:link:focus,.css-yidnqd-InlineLink:visited:focus{border-bottom-color:currentcolor;border-bottom-width:2px;color:#B80000;}@supports (text-underline-offset:0.25em){.css-yidnqd-InlineLink:link,.css-yidnqd-InlineLink:visited{border-bottom:none;-webkit-text-decoration:underline #BABABA;text-decoration:underline #BABABA;-webkit-text-decoration-thickness:1px;text-decoration-thickness:1px;-webkit-text-decoration-skip-ink:none;text-decoration-skip-ink:none;text-underline-offset:0.25em;}.css-yidnqd-InlineLink:link:hover,.css-yidnqd-InlineLink:visited:hover,.css-yidnqd-InlineLink:link:focus,.css-yidnqd-InlineLink:visited:focus{-webkit-text-decoration-color:currentcolor;text-decoration-color:currentcolor;-webkit-text-decoration-thickness:2px;text-decoration-thickness:2px;color:#B80000;}}and listen/speak via an intercom-like facility.

Battle in EU over burgers+sausages being replaced by vegan “discs” and “fingers”

A battle over whether vegan burgers and sausages should be replaced by vegan “discs” and “fingers” is intensifying in the EU ahead of a vote in European Parliament on a contentious food labelling amendment next week.

On Wednesday, MEPs are scheduled to vote on whether to limit the use of words such as “steak”, “sausage”, “escalope” and “burger” on labels for products containing meat. The move has been backed by the region’s meat and livestock industry and has won the support of the European Parliament’s agricultural committee.

The vote comes as vegan alternatives are growing popular, boosted by new entrants into the plant-based substitutes market which mimic the look, taste, and mouth feel of real meat. The pandemic has hit meat processors and also provided a tailwind for sales of plant-based products, which have jumped 73 per cent in Europe over the past five years, according to Euromonitor, the consumer data group.

MEPs and campaigners said the vote hung in the balance. If approved by a majority of MEPs, the amendment, which is part of a reform of the EU common agricultural policy, will be discussed as part of further negotiations between the parliament and EU governments. The measure will only make it to law if it is agreed by both institutions. 

Livestock and meat industries claim that the use of meat related terms and names confuse consumers. “If we do not protect [meat products], we risk misleading consumers and allowing the hijacking of the work done by the livestock chain over the years to develop their renowned products,” said Copa Cogeca, the association of European farmers and co-operatives.

Opponents argue a change in labelling would also confuse consumers. Elena Walden, Europe policy manager of plant-based meat advocacy group Good Food Institute, described a potential EU ban as “patronising” for consumers, while Jytte Guteland, a Swedish MEP, said names such as “veggie burger” had been around for a long time. “It’s important that we help consumers to make climate friendly choices and for them to eat less meat,” she said.

Some meat processors and distributors have started offering their own vegan products or forming partnerships with alternative protein producers, triggering the ire of livestock farmers.

“What is vegan meat?” said Antonio Tavares, a Portuguese pork farmer who also grows grains and vegetables. “I’m not saying vegan products are bad for our health, but it’s [a] completely different [product]. A burger, it’s meat.”

EU member states already have the power to issue their own food labelling laws in a bid to prevent consumers from being misled. Earlier this year, France passed a similar legislation to ban the use of meat nomenclature for vegetarian and vegan substitutes. 

However, the Dutch government last year announced that plant-based meat producers can use terms such as “chicken” as long as it was clearly marked that it was a vegan or vegetarian product. 

Europe has already clamped down on dairy substitutes when the European Court of Justice in 2017 banned the use of terms such as “milk”, “butter”, and “yoghurt” for marketing non-animal products.

A move to stop plant-based meat producers from using meat-related nomenclature is also spreading across in the US, with individual states considering or passing legislation. However vegan food companies and campaigners are launching legal challenges against such laws in the federal courts. 

Coursera’s co-founder thinks Zoom doesn't work for learning

Daphne Koller changed the world of online education once when she co-founded Coursera. Now she hopes to do it again.

The pandemic has forced universities to host most of their classes online, recreating the in-person experience for students on the web. Koller believes she can challenge Zoom with a video platform tailor-made for education called Engageli.

“I think we’re finally at a place where, because of this pandemic, we need to build the technology to support the learning experience that frankly, most people would have benefited from even without the pandemic,” she told Protocol.

Get Protocol straight into your inbox

With Coursera, started alongside fellow Stanford professor Andrew Ng, Koller pitched universities on the idea of opening up their courses to an online audience that wasn’t enrolled in the school, creating massive open online courses or MOOCs. Elite universities bought into the vision. But now, she thinks, enrolled students could use some help.

While many classes have moved straight to Zoom, Engageli is one of a handful of “Zoom for X” startups that are building alternative platforms catering to a specialized purpose. Unlike a Zoom call that ends up looking like a round of “Hollywood Squares,” Engageli’s platform is designed to let students “sit” at smaller virtual tables with groups of students, where they can chat or work on assignments while following along with the instructor. There’s a built-in note-taking tool that syncs with the timestamp in the recording, plus different ways to do quizzes, polls and chats inside the window.

Professors will be able to see indicators of a student’s engagement with the software to see who is paying attention and taking notes or responding to quizzes, versus students who maybe are distracted by something else. Engageli considered technologies like eye tracking, which has been a privacy concern, to monitor students, but ultimately decided against it. Instead, the engagement metrics that she decided to track are focused on how much students are actually interacting with the platform and each other.

“There’s so many questions about education that we’ve never been able to get rigorous answers to, because we just haven’t been able to perform an experiment where we can measure outcomes,” Koller told Protocol. “Now, we can actually do that.”

Investors have already bought into the vision; now Engageli needs to sell it to the universities. The company has been operating an invitation-only pilot program and is emerging from stealth on Wednesday, having raised a $14.5 million seed round. Koller co-founded the company alongside her husband and serial entrepreneur Dan Avida, who is Engageli’s CEO; Stanford professor emeritus Serge Plotkin; and 2U executive Jamie Nacht Farrell. Koller will remain the CEO of Insitro, a company focused on drug development, and serve as a board member to Engageli in her spare time. But as someone who helped bring online classes to the masses, and now has her own children in high school going through the process, she’s ready to again build a better online education experience.

Protocol spoke to Koller about building Engageli and Coursera, how technology can improve access to education and why Zoom falls short when it comes to online learning.

The following Q&A has been edited for length and clarity.

You’re already the founder and CEO of a startup working on big hairy problems around drug discovery and medicine. Why start another company, especially in the middle of a pandemic?

In some ways, I can tie it back to my early Stanford experience where, as a big research university, people are mostly rewarded on the research, and I’ve always had this soft spot for teaching and always had these side passion projects there. I think learning is so important, and I wanted to help make that better, so we carved out the time from my schedule to do that. My current job doesn’t really involve much of that, so I think I’ve been feeling [a] lack of that.

And then the pandemic, if anything, made this a particularly pointed moment at which to do this. Because even within my own house, the learning experience that my high school kids were getting as we were all forced to switch to a Zoom-based education in the spring … honestly, it wasn’t a great learning experience. They’re in a great school with relatively small classes and even so, it was clear that we were trying to take the teaching experience and push it onto a platform that was never intended for teaching. It was for video conferencing for business people.

That really created the impetus and a time to think differently and bring back ideas and lessons that I had in my Stanford days and really build it into a platform that fits the purpose for education.

With your new company, Engageli, how are you doing things differently than using Zoom? What needed to change that Zoom was not enough to recreate that classroom experience?

That’s a great question, and the platform has several pillars that are critical and in many ways are an improvement over what face-to-face instruction was.

One is the focus on engagement of the students with the material – both techniques for creating that engagement by different mechanisms, but also importantly measuring the engagement and reflecting that to the instructor so that the instructor can see who is paying attention, who is actually noticing, and really get a pulse for their class so that they can call on the people who are not paying attention or change their teaching if the majority of the classes isn’t following. That sort of pulse is really critical.

The second one is the focus on peer-to-peer interaction. There are these breakout rooms in Zoom, but they feel very appropriate for, again, a conferencing system and not for the kinds of very dynamic interactions. Like in my daughter’s classrooms, one of the things that they used to have was these table groups where you get to sit with a few of your friends, chit chat with them quietly, while also listening to the instructor and do these activities, as a small group. Organic interaction is not something that you get using a traditional breakout group model.

A third one, which to me is something I’ve really been passionate about because of my scientific work, is the collection of data that relates to learning activities, learning outcomes, that allows us to learn from what we see works and what doesn’t, and then adapt our teaching accordingly. So this is an incredibly instrumented platform that allows us to collect data that frankly, in traditional educational settings, we just were never able to collect.

What’s an example of that data?

There’s all these engagement tools on the platform — like upvot[ing] something, you can ask a question or there’s polls, there’s exercises that are integrated into the learning experience — all of those are tracked and stored. You can then start to ask really important questions like, what kinds of engagements are most predictive of ultimate success? What happens if I add an intervention? What happens if the instructor actually calls on a student? Does that actually influence it?

You can actually start to evaluate different teaching strategies, and see which of them has an effect and which one doesn’t. To give you another example, we all believe in peer-to-peer interaction and study groups, but the amount of data on what’s the appropriate formation of a study group is — is it all people at the same level? People who are on different levels to help teach the ones who are a little further behind? — you don’t really know. We’ve never been able to measure this. There’s so many questions about education that we’ve never been able to get rigorous answers to, because we just haven’t been able to perform an experiment where we can measure outcomes. Now, we can actually do that.

Engageli cofounder Daphne Koller thinks her new startup can help teachers better understand how engaged their students are.Image: Engageli

When you are talking to universities and teachers, what do they seem to be missing or desiring the most out of rebuilding their classrooms online these days?

First of all, just the ability to engage with the students and see are they paying attention. Regardless of whether they’re there physically, are they there mentally?

The data analytics is a huge part of it. Even prior to this whole pandemic, it was something that a lot of instructors wanted to get a sense of. So if I’m doing my best to teach students, is it working? There’s really no way to test that other than at the end of the semester, and you look at the final exam and say how on earth did I not know that 70% of the class doesn’t get this basic concept?

And then the last thing, which I think is really important, is that education needs to be something that everyone has access to. That was problematic before, but it’s even more problematic now. There are people who lack basic bandwidth, the infrastructure of the computers, and you [are] really shoving a lot of people out of educational experience because of those requirements. And one of the focus areas of Engageli is the ability to use relatively simple, cheap mobile devices, that we all have, and really create an educational experience that is much more accessible.

This isn’t the first time you’ve been working on a startup that’s essentially bringing education online. What was an early lesson you learned from building Coursera that you found now is applicable again here in 2020?

One of the lessons that I learned at Coursera is really just the importance of engagement. At Coursera we had students from all of the world taking classes, and it was covered extensively in the media, the lower completion rates. We said at the time, and I think it still holds true, a lot of students didn’t complete the courses because they weren’t in it to complete the courses. They came in mostly to maybe learn a few modules and walk away knowing more than they had learned before, kind of like if you pick up a book and read three or four chapters.

But then we did have a lot of students who wanted to complete the courses, but just weren’t able to sustain enough engagement to get all the way through. As we started to build out some of the paid programs on Coursera, we found how difficult it was to add synchronous components of interaction between the instructor and the students and how much that made a huge difference in the completion rate. So, really, I think what I learned from that is there are different kinds of learning experiences.

A question I’m asking all edtech startups right now is how do you know you’re building something that will outlast this pandemic versus something that is helpful in this moment. Once schools go back, will demand for Engageli be gone?

I don’t know if in K-12 people are going to do online Zoom calls in large numbers. We can debate whether that’s a good thing or a bad thing, but I think a lot of kids as well as parents really want the kids to be out of the house, in a world where they can play with other kids, and so on and so forth. So will that have a large online component? I don’t know.

For university students, especially for the ones who have other constraints on their life, I think there’s going to be a significant portion that want to take either some or all of their classes in an online format. That was already a demand, and that trend was already happening, this really just accelerated it by five or 10 years.

If you’re holding down a job and trying to raise a kid at the same time, you don’t have that utopian residential experience where you’re in a green lawn sitting with other people talking about philosophy or science, walking from one beautiful building to another. You’re probably in a community college or state university, driving frantically from your job to take a class and then driving back and then finding time at night to work on your assignments. That’s what college is like for a lot of people.

It sounds like we’re almost at the point where technology is actually catching up to what the reality of education is like versus the other way around where technology is what’s driving the change.

I think we’re finally at a place where, because of this pandemic, we need to build the technology to support the learning experience that, frankly, most people would have benefited from even without the pandemic.

It’s not just about whether you can take the courses through your cell phone: It’s whether your learning fits inside your life, which I think pushes a lot of people out of the education system. That’s why there’s a large group of people who have some college with no degree, because their life just couldn’t accommodate it.

Doing Old Things Better vs. Doing Brand New Things

New technologies enable activities that fall into one of two categories: 1) doing things you could already do but can now do better because they are faster, cheaper, easier, higher quality, etc. 2) doing brand new things that you simply couldn’t do before. Early in the development of new technologies, the first category tends to get more attention, but it’s the second that ends up having more impact on the world.

Doing old things better tends to get more attention early on because it’s easier to imagine what to build. Early films were shot like plays — they were effectively plays with a better distribution model — until filmmakers realized that movies had their own visual grammar. The early electrical grid delivered light better than gas and candles. It took decades before we got an electricity “app store” — a rich ecosystem of appliances that connected to the grid. The early web was mostly digital adaptations of pre-internet things like letter writing and mail-order commerce. It wasn’t until the 2000s that entrepreneurs started exploring “internet native” ideas like social networking, crowdfunding, cryptocurrency, crowdsourced knowledge bases, and so on.

The most common mistake people make when evaluating new technologies is to focus too much on the “doing old things better” category. For example, when evaluating the potential of blockchains, people sometimes focus on things like cheaper and faster global payments, which are important and necessary but only the beginning. What’s even more exciting are the new things you simply couldn’t create before, like internet services that are owned and operated by their users instead of by companies. Another example is business productivity apps architected as web services. Early products like Salesforce were easier to access and cheaper to maintain than their on-premise counterparts. Modern productivity apps like Google Docs, Figma, and Slack focus on things you simply couldn’t do before, like real-time collaboration and deep integrations with other apps.

Entrepreneurs who create products in the “brand new things” category usually spend many years deeply immersed in the underlying technology before they have their key insights. The products they create often start out looking toy-like, strange, unserious, expensive, and sometimes even dangerous. Over time, the products steadily improve and the world gradually embraces them. 

It can take decades for this process to play out. It’s clear that we are early in the development of emerging technologies like cryptocurrencies, machine learning, and virtual reality. It is also possible we are still early in the development of more established technologies like mobile devices, cloud hosting, social networks, and perhaps even the internet itself. If so, new categories of native products built on top of these technologies will continue to be invented in the coming years.

Ask HN: State of the art architecture for paired comparison prediction?

Ask HN: State of the art architecture for paired comparison prediction?
1 point by morelandjs 14 minutes ago | hide | past | favorite | discuss
I want to learn more about the domain of creating time series predictions for paired comparison outcomes, e.g. taking historical box scores and predicting future box scores.

One of the primary challenges is that such data includes both numeric features (e.g. box stats) and contextual labels (e.g. teams), making each observation both relative and coupled.

Perhaps the most natural representation of such data is a time-dependent graph, where the comparisons represent edges between two nodes (competitors). Unfortunately, this kind of graph structure appears to preclude naive applications of ANNs developed for tabular and sequence data (e.g. LSTMs, transformers, etc).

Does anyone know of any good references that demonstrate how to account for the graph structure of such problems? Is it a terrible idea to essentially one-hot encode the adjacency matrix and provide it as an input to an LSTM or Transformer architecture? Are there more efficient ways to encode paired comparisons?

I tried something like a neural network version of Elo with more features (i.e. RNN), but I suspect I was suffering from exploding/vanishing gradients since I was chaining the same hidden layer update step for thousands of games.

Any advice or direction here would be much appreciated!

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Ed Benguiat, a Master of Typography, Is Dead at 92

Ed Benguiat, a celebrated graphic designer known for his expertise in typefaces — including the one you see at the top of the print and web editions of this newspaper — started his design career in a not-so-celebrated post at a movie magazine publisher.

“My job was to be a cleavage retoucher,” he recalled in a video interview with the Type Directors Club. “My job was to take it out — take the cleavage out, remove it.”

It was the years after World War II, an era of the restrictive Hays Code in the movies.

“I was very good with an airbrush and buying doilies in the 5 and 10,” he said, strategically placed doilies being key to the cleavage removal process.

Mr. Benguiat went on to more sophisticated work. He became one of the go-to designers of the second half of the last century, especially in matters of typography. His hand was behind more than 600 typefaces, several of which bear his name (which is pronounced ben-GAT). The Telegraph of Britain, in a 2016 article about him prompted by the striking use of one of his fonts (ITC Benguiat) in the title sequence of the hit Netflix series “Stranger Things,” called him “one of the type industry’s greats.”

Mr. Benguiat died on Thursday at his home in Cliffside Park, N.J. He was 92.

His wife of 38 years, Elisa (Halperin) Benguiat, confirmed the death.

Mr. Benguiat was an important figure in the design world for a number of reasons. According to his citation in the Art Directors Club Hall of Fame, where he was inducted in 2000, he helped establish the International Typeface Corporation, the first independent licensing company for type designers, and became its vice president. He also taught for almost 50 years at the School of Visual Arts in Manhattan.

But it was his painstaking work designing new typefaces and modifying existing ones that made him a revered figure in the business, and that reached the public eye, although the public rarely knew his role. He designed logotypes for companies including Ford and AT&T and for Esquire, Look, McCall’s and other publications. His typefaces were seen in movies including “Super Fly” (1972) and “Planet of the Apes” (1968).

Credit…20th Century Fox

Mr. Benguiat understood the intricacies of a typeface in a way that today’s computer users, with countless fonts at their disposal, generally do not. He knew that a successful design wasn’t merely in the shaping of individual letters; it was in things like the spacing between those letters. And he knew that what looks good on a computer screen might not work when blown up to the size of a marquee or a billboard.

“At three feet high, the serif of a face like Bodoni is going to be two inches thick,” he told Macworld in 2001, referring to a popular typeface. “Someone has to fix it. I get called to do that.”

One such “fixer” assignment, in 1967, was for The New York Times. Louis Silverstein, the paper’s promotion art director at the time, was given the task of revisiting the nameplate, which had been tweaked over the previous century but remained a distinctive calling card. Mr. Silverstein tweaked it anew.

“To strengthen the logo, I redrew it,” he wrote later, “making the thicks thicker and the thins thinner.”

“I drew the new logo on tracing paper,” he added, “and hired Ed Benguiat to do the actual ink drawing. Ed was perhaps the most accomplished letterer in the country.”


Credit…Milton Glaser Design Study Center and Archives/Visual Arts Foundation

One result of that redesign was the disappearance of the period that for decades had come after the word “Times” in the logo. (Some readers mourned its loss. “No tittle in your title?” one wrote.) In the Macworld interview, Mr. Benguiat recalled the Times logo assignment this way:

“Lou Silverstein was the art director. His thought was, ‘Change it.’ My thought was, ‘OK, we’ll change it — but if we change it, nobody will recognize it.’ So all I did was take it and fix it.”

Ephram Edward Benguiat was born on Oct. 27, 1927, in Brooklyn. His mother, Rose (Nahum) Benguiat, was a driver for the Red Cross, and his father, Jack, was design director at Bloomingdale’s; Mr. Benguiat often spoke of his childhood fascination with his father’s pens and paintbrushes.

Some articles about Mr. Benguiat over the years said that one of his first efforts at tweaking type was when he forged a birth certificate to make himself appear old enough to join the Army during World War II, but in a 2017 talk at the Type Directors Club, he corrected that; it was his father who did the forging, he said. It was good enough to get him into the Army Air Forces, and during the war he was stationed in Italy, serving first as a radio operator on a bomber and then doing photo reconnaissance.


Credit…David Zalubowski/Associated Press

Mr. Benguiat had been playing the drums since his father bought him a drum set at age 10, and under the name Eddie (or sometimes E.D.) Benart, he played with various jazz ensembles, including those of Woody Herman and Stan Kenton. But the work lost its allure.

“One day I went to the musician’s union to pay dues and I saw all these old people who were playing bar mitzvahs and Greek weddings,” he said. “It occurred to me that one day that’s going to be me, so I decided to become an illustrator.”

An establishment near one of the clubs where he played beckoned.

“There was a sign on Fifth Avenue; it said, ‘Draw me,’” he said in the 2017 talk. “So I went upstairs and I registered.”

It was the Workshop School of Advertising Art, where he studied layout, design, typography and calligraphy. The cleavage-covering job, he said, developed into something more by happenstance.

“The lettering man was gone, and something was missing,” he recalled. “I said, ‘I can do it,’ and I did it, and that’s what started the ball rolling.”

Eventually he had enough skills to be hired, in 1953, by Esquire magazine, and in 1962 he joined Photo-Lettering Inc., a typesetting company, as typographic design director. He developed some 400 typefaces there.

In 1971 he joined the newly established International Typeface, making a quick impact there by retooling the typeface Souvenir. His revised version became immensely popular.


Credit…Milton Glaser Design Study Center and Archives/Visual Arts Foundation

As for his own typefaces, Mr. Benguiat said that developing one from scratch could take him a year or more. He wasn’t hostile to computers when they arrived and changed how graphic design was done, but he maintained that good design started with a good hand.

“If you can’t draw a shape that’s pleasing on a piece of paper,” he told Macworld, “how the hell are you going to do it on the screen?”

In a 1989 interview with The Times, he put it another way. “The most beautiful thing in the world,” he said, “is a blank piece of paper.”

In addition to his wife, Mr. Benguiat is survived by a granddaughter and two great-grandchildren. A son from an earlier marriage, Jon, died in January.

In the Type Directors Club video, Mr. Benguiat said he saw a connection between his early career as a musician and his later one.

“Music is placing sounds, to me, in their proper order so they’re pleasing to the ear,” he said. “That’s all. What is graphic design? Placing things in their proper order so they’re pleasing to the eye.”