Could LLMs Revolutionize Drug and Material Design?
These researchers have developed an innovative system that augments an LLM with graph-based AI models, designed specifically to handle molecular structures
A new method is changing the way we think about molecule design, bringing us closer to the possibility of using large language models (LLMs) to streamline the creation of new medicines and materials. Imagine asking, in plain language, for a molecule with specific properties, and receiving a comprehensive plan on how to synthesize it. This futuristic vision is now within reach, thanks to a collaboration between researchers from MIT and the MIT-IBM Watson AI Lab.
A New era in molecular discovery
Traditionally, discovering the right molecules for medicines and materials has been a slow and resource-intensive process. It often involves the use of vast computational power and months of painstaking work to explore the nearly infinite pool of potential molecular candidates. However, this new method, blending LLMs with other machine-learning models known as graph-based models, offers a promising solution to speed up this process.
These researchers have developed an innovative system that augments an LLM with graph-based AI models, designed specifically to handle molecular structures. The approach allows users to input natural language queries specifying the desired molecular properties, and in return, the system provides not only a molecular design but also a step-by-step synthesis plan.
LLMs and graph models
LLMs like ChatGPT have revolutionized the way we interact with text, but they face challenges when it comes to molecular design. Molecules are graph structures—composed of atoms and bonds—which makes them fundamentally different from text. LLMs typically process text as a sequence of words, but molecules do not follow a linear structure. This discrepancy has made it difficult for LLMs to understand and predict molecular configurations in the same way they handle sentences.
To bridge this gap, MIT’s researchers created Llamole—a system that uses LLMs to interpret user queries and then switches between different graph-based AI modules to generate molecular structures, explain their rationale, and devise a synthesis strategy. The system combines the power of text, graphs, and synthesis steps into a unified workflow.
As a result, this multimodal approach drastically improves performance. Llamole was able to generate molecules that were far better at meeting user specifications and more likely to have a viable synthesis plan, increasing the success rate from 5 percent to 35 percent.
Llamole’s success lies in its unique ability to seamlessly combine language processing with graph-based molecular modeling. For example, if a user requests a molecule with specific traits—such as one that can penetrate the blood-brain barrier and inhibit HIV—the LLM interprets the plain-language request and switches to a graph module to generate the appropriate molecular structure.
This switch occurs through the use of a new type of trigger token, allowing the LLM to activate specific modules as needed. The process unfolds in stages: the LLM first predicts the molecular structure, then uses a graph neural network to encode the structure, and finally, a retrosynthetic module predicts the necessary steps to synthesize the molecule. The seamless flow between these stages ensures that the LLM maintains an understanding of what each module does, further enhancing its predictive accuracy.
“The beauty of this is that everything the LLM generates before activating a particular module gets fed into that module itself. The module is learning to operate in a way that is consistent with what came before,” says Michael Sun, an MIT graduate student and co-author of the study.
Simplicity meets precision
One of the most striking aspects of this new method is its ability to generate simpler, more cost-effective molecular structures. In tests, Llamole outperformed other LLM-based methods and achieved a notable 35 percent success rate in retrosynthetic planning, up from a mere 5 percent with traditional approaches. “On their own, LLMs struggle to figure out how to synthesize molecules because it requires a lot of multistep planning. Our method can generate better molecular structures that are also easier to synthesize,” says Gang Liu, the study’s lead author.
By designing molecules with simpler structures and more accessible building blocks, Llamole could significantly reduce the time and cost involved in developing new compounds.
The road ahead
Though Llamole’s current capabilities are impressive, there is still work to be done. The researchers built two custom datasets to train Llamole, but these datasets focus on only 10 molecular properties. Moving forward, they hope to expand Llamole’s capabilities to design molecules based on a broader range of properties and improve the system’s retrosynthetic planning success rate.
In the long run, the team envisions Llamole serving as a foundation for broader applications beyond molecular design. “Llamole demonstrates the feasibility of using large language models as an interface to complex data beyond textual description, and we anticipate them to be a foundation that interacts with other AI algorithms to solve any graph problems,” says Jie Chen, a senior researcher at MIT-IBM Watson AI Lab.
With further refinements, Llamole could revolutionize fields from pharmaceuticals to material science, offering a glimpse into the future of AI-driven innovation in molecular discovery.
EP Staff is the editorial team at EdPublica, an independent media organisation focused on science, education, environment and public policy. The team produces evidence-based news, features, explainers and analysis on issues that shape society and everyday life.
What is the Mysterious ‘Egoviruses’ in the Guts of 180 Million People
Scientists have discovered egoviruses, a previously unknown group of gut viruses potentially present in 180 million people worldwide. Their genetic connection to poxviruses could offer new clues about viral evolution.
A recent study identified a previously unknown group of gut-associated viruses called egoviruses, with close evolutionary links to poxviruses. Representational image. Image credit: Merlin Lightpainting/ Pexels
Millions of people could be carrying a previously unknown group of disease-causing viruses in their digestive systems without ever knowing it. Scientists have now identified these microbes, named ‘egoviruses’, in human and animal gut samples, uncovering a surprising connection to one of the most important families of disease-causing viruses.
The discovery is drawing attention because egoviruses are closely related to poxviruses, the family that includes the viruses responsible for smallpox and mpox. Researchers estimate that these newly identified viruses could be present in around 180 million people worldwide. The findings offer a new perspective on the evolution of viruses and the largely unexplored microbial world inside the human digestive system.
Hidden Viral Population in the Human Gut
The discovery emerged from an extensive examination of genetic material collected from human and animal digestive systems. Rather than isolating individual viruses in laboratories, researchers used metagenomics, a technique that allows scientists to identify organisms by analysing genetic material in biological samples.
Scientific analysis of egovirus genomes reveals their evolutionary relationship with poxviruses and genetic similarities with viruses found in single-celled organisms living in animal digestive systems. Source: Nature research study.
The research team, led by microbial ecologist Tom Delmont of the French National Centre for Scientific Research, identified egoviruses across samples from several countries. These included Tanzania, Nepal, Fiji, the United States and parts of Europe.
The analysis revealed six distinct groups associated with humans. One of these groups was found across multiple continents, suggesting that the viruses are not confined to a particular population or geographical region. The researchers also detected egoviruses in buffaloes, goats, chickens, deer and pigs. Their presence across different animal species points to a much wider distribution than scientists had previously recognised.
Do These Viruses Actually Infect Humans?
If egoviruses are found in human gut samples, does that mean they infect human cells? Not necessarily. The researchers suspect that these viruses may target single-celled organisms living in the digestive tract rather than human cells themselves. Genetic comparisons revealed similarities between egovirus genes and those of two groups of single-celled organisms, Parabasalia and Trichostomatia.
Some members of these groups live in animal digestive systems. Trichostomatia, for example, includes microorganisms found in the digestive tracts of ruminants, where they are involved in breaking down food. This could mean that the human gut contains viruses that interact with its microbial inhabitants rather than directly attacking human tissue. However, researchers have yet to confirm their suspected hosts through direct experiments. There is also no evidence from this study that egoviruses cause illness in humans.
Surprising Link to Smallpox and Mpox Viruses
Perhaps the most intriguing aspect of the discovery lies in the evolutionary history of egoviruses. The researchers found that these viruses are the closest known relatives of poxviruses among the groups they examined. Poxviruses are large DNA viruses that can infect humans and animals. Smallpox, eradicated in 1980, and mpox belong to this family.
The genetic relationship has led researchers to investigate whether the ancestors of poxviruses once infected single-celled organisms before evolving the ability to infect animal cells. One possibility is that an ancestral virus initially lived among microorganisms in the digestive systems of early animals. Over millions of years, evolutionary changes may have enabled some descendants to infect animal cells directly.
This remains a hypothesis, rather than a confirmed account of how poxviruses originated. Nevertheless, egoviruses offer scientists a previously unknown group of viruses through which to investigate that evolutionary transition.
What Scientists Still Need to Find Out
The discovery adds another layer to the growing understanding of the human gut, which hosts a complex community of bacteria, fungi, viruses and other microorganisms. But identifying a virus is only the beginning. Scientists still need to establish which organisms egoviruses infect, how they behave inside the gut and whether they influence the wider microbial community.
For now, the discovery of a virus group potentially carried by 180 million people raises a fascinating possibility, a large part of the human body’s viral diversity may have remained hidden simply because scientists had not yet known where to look.
India’s medical students show interest in research, but mentorship and institutional support remain major hurdles
A new review highlights the barriers facing Indian medical students pursuing research, including inadequate mentorship, limited funding and insufficient institutional support, despite strong student interest.
Medical students participate in a laboratory practical session, highlighting the importance of hands-on scientific training and research exposure in medical education.Image credit:Pixels
Indian medical students recognise the importance of research in strengthening healthcare, but many struggle to translate that interest into actual research projects and publications. Limited faculty mentorship, inadequate funding, insufficient research training and the pressure of academic workloads continue to restrict undergraduate research participation, according to a review published in Frontiers in Medicine on November 13, 2026.
The findings point to a gap between students’ willingness to pursue research and the institutional support available to help them do so.
Research interest does not translate into participation
A 2024 cross-sectional study involving 606 Indian medical students found that awareness of research opportunities did not necessarily translate into completed projects.
Although 77.23% of respondents knew about the Indian Council of Medical Research’s (ICMR) Short-Term Studentship (STS) programme, only 16.34% had applied. Just 6.4% completed their research within the prescribed period, while only 1.98% published their ICMR-related findings.
The study also examined students’ access to academic guidance. While 65.35% approached faculty members for assistance, only 46.53% found them helpful. Just 33.17% received assistance from senior students.
The findings suggest that awareness campaigns alone may not be sufficient to develop a research culture without accessible mentorship and sustained institutional support.
The 2026 review also highlights findings from individual medical colleges. At ESIC Medical College, Chennai, a study involving 241 students found that only 8.3% actively participated in research, despite 60% recognising its importance in clinical care.
At Pushpagiri Institute of Medical Sciences and Research Centre in Kerala, a study of 160 undergraduate students and interns found that 96.9% recognised the value of research, but only 12.5% had participated in ICMR-STS.
These institution-specific findings illustrate the difference between research awareness and actual engagement.
A wider institutional research gap
The challenge extends beyond undergraduate classrooms. A study examining research output across 579 Indian medical institutions, cited in a 2026 research paper, found that 57.3% had no publications.
The finding raises questions about the research environment in medical institutions and the opportunities available to students who want to participate in scientific investigation.
However, some institutions have demonstrated that structured mentorship can improve participation.
A study published in BMC Research Notes in April 2026 documented the experience of JIGNYASA, a student-led undergraduate research committee established in February 2020 at a government medical college and tertiary-care hospital in western India.
Between February 2020 and September 2025, the initiative recorded 228 student-initiated research projects, 34 grant approvals and 59 publications.
A medical student examining a report.Image credit:Pexels
The model relied on senior-student mentorship, continuous faculty guidance, research-promotion activities and regular feedback. The authors presented it as a potentially replicable approach for institutions seeking to strengthen undergraduate research without substantial additional funding.
National programmes need stronger implementation
The Frontiers review examines three initiatives: ICMR’s Short-Term Studentship programme, the National Medical Commission’s Community Health Programme and the Basic Course in Biomedical Research.
The STS programme provides undergraduate medical and dental students an opportunity to undertake short-term research under faculty supervision. The Community Health Programme introduces students to field-based health research, while the Basic Course in Biomedical Research provides structured training in research methodology.
However, the review identifies limitations, including inadequate research grants, uneven access to mentors, limited hands-on training and insufficient support for publishing findings.
It argues that medical colleges need to provide dedicated research time, accessible funding, trained faculty mentors and stronger institutional research infrastructure.
For India, developing clinician-researchers is important not only for improving medical education but also for strengthening the country’s capacity to generate evidence relevant to its own healthcare needs.
The evidence suggests that the problem is not simply a lack of interest among medical students. It is the absence of consistent opportunities and institutional structures that can turn that interest into meaningful scientific contributions.
Six months on, Ebola outbreak in DR Congo remains uncontrolled, with over 5,500 deaths
Six months after its declaration, the Ebola outbreak in DR Congo remains uncontrolled, with over 5,500 deaths. Armed conflict, gaps in disease surveillance and limited healthcare access continue to hamper containment efforts as the Bundibugyo virus spreads across eight provinces.
A participant in the eastern DR Congo receives a vaccine as part of a trial to tackle the Ebola Bundibugyo virus.Image credit: WHO
Six months after the Democratic Republic of the Congo (DRC) declared an Ebola outbreak, the disease remains out of control, with more than 11,300 confirmed cases and 5,500 deaths reported in the country. Despite an expanded international response, the outbreak continues to spread amid armed conflict, displacement and weaknesses in healthcare delivery.
The Africa Centres for Disease Control and Prevention (Africa CDC) warned on November 15 that the outbreak was still not under control. Its Director-General, Jean Kaseya, cautioned that unless transmission is interrupted, the epidemic could become the world’s worst Ebola outbreak.
According to the European Centre for Disease Prevention and Control (ECDC), the DRC had recorded 11,376 confirmed infections and 5,568 deaths as of November 15, based on data up to November 14. Eight provinces have been affected, with Ituri reporting the highest burden. The outbreak is caused by the Bundibugyo virus, a relatively rare species of Ebola virus.
Delayed detection is driving transmission
One of the biggest obstacles to containment is the inability to identify infected people and their contacts early. According to Africa CDC, more than 80% of new infections are being detected among people who were not previously listed as contacts of confirmed patients.
Contact tracing is central to Ebola control. Once an infection is confirmed, health workers identify people who may have been exposed and monitor them for symptoms during the 21-day incubation period. Failure to identify these contacts creates opportunities for further transmission.
More than 65% of deaths have occurred in communities rather than healthcare facilities, indicating that many patients are being identified too late to receive timely treatment. The World Health Organization (WHO) reported that children account for more than a quarter of confirmed infections and approximately one-third of deaths. Seven in 10 children under five who contract the disease have died.
The outbreak has expanded to 67 health zones across eight provinces, with an estimated 400–500 new cases being reported each week, according to the WHO’s November 14 update.
Conflict and insecurity disrupt response
Armed conflict and population displacement have further complicated containment. Insecurity has restricted access to affected communities, disrupted healthcare services and made surveillance and contact tracing difficult.
The WHO reported more than 80 attacks on healthcare since the outbreak began. An armed attack in Cantine-Aloya killed 66 civilians, including three health workers, while three healthcare facilities were looted. More than 1,000 contacts were subsequently lost to follow-up, and three confirmed patients could not be accounted for.
The response has also been affected by delayed payments to frontline health workers. Africa CDC has warned that inadequate resources, insufficiently trained personnel and operational difficulties are weakening efforts to contain the disease.
To help fight the spread of Ebola, UNICEF is supporting District Health Educators (DHE) in high-risk districts to conduct outreach activities with different groups, including corporate and business sectors, educational facilities, prisons, hostels and guest-houses. Photo: UNICEF/UNI790701/Bamulanzeki
Healthcare capacity expands, but gaps persist
There has been progress in strengthening the response. According to the WHO, diagnostic capacity has increased from two laboratories to more than two dozen, while nearly 2,000 treatment beds have been made available across 104 treatment and transit centres.
More than 17,000 community health workers and 46,000 community actors have received training to support surveillance, contact tracing and public awareness. Nearly 90% of registered deaths are now being managed through safe and dignified burials, reducing the risk of transmission during funerals.
However, these improvements have not translated into nationwide containment. Several initial outbreak centres, including Butembo, Mongbwalu and Nyankunde, have brought transmission under control, but new infections continue to emerge elsewhere.
The challenge of containing Bundibugyo virus
The virus itself presents another challenge. Bundibugyo is distinct from the Zaire ebolavirus responsible for the 2014–2016 West African epidemic. Vaccines developed against the Zaire species cannot be assumed to protect against Bundibugyo virus. There is currently no licensed vaccine specifically approved against this strain.
The international response therefore continues to depend on early diagnosis, isolation, supportive treatment, infection prevention and safe burial practices, alongside research into candidate vaccines and treatments.
The WHO has called for sustained, flexible funding through 2027 and improved access to communities affected by conflict. The outbreak has also prompted stronger cross-border surveillance, following infections reported in Kenya and Nigeria.
Six months into the crisis, the challenge is no longer simply mobilising an emergency response. It is ensuring that diagnostic services, trained personnel and treatment reach communities where transmission continues. Without sustained funding, improved security and effective contact tracing, the outbreak could remain a regional public health threat.