Lifeboat Foundation: Safeguarding Humanity
Toggle light / dark theme

Blog

Category: bioengineering – Page 6

PlasmidGPT: A generative framework for plasmid analysis and generation

By training an AI model using 153,208 plasmids from Addgene, Shao et al.’s PlasmidGPT can annotate and classify existing plasmids as well as generate new functional plasmid sequences from DNA “prompts”. While existing plasmid design tools currently surpass PlasmidGPT in sophistication, future architecture and training data augmentations may allow us to automate much of the plasmid design process. I could certainly see this playing a role in high-throughput biological screening methods.

Assembly standards facilitate the construction of functional plasmids (13). Collections such SEVA (14, 15) and CIDAR MoClo (16) include ready-to-use constructs and genetic parts that can be easily assembled, making them valuable tools for microorganism bioengineering and genome editing. Tools such as Cello (17, 18) can design genetic constructs with a high success rate for specific functions, such as computation, across diverse organisms. iBioSim 3 enables the design and modeling of genetic circuits that extend beyond logic circuits (19). However, there is still no computational method capable of harnessing the existing collection of plasmid sequences for designing the full spectrum of plasmids, such as those for mammalian expression, bacterial expression, and gateway vectors. Consequently, for many applications, plasmid DNA design remains a labor-intense process that requires manual inspection, annotation, and the combination of functional sequences.

Recently, generative models such generative pretrained transformers (GPTs) (20) have demonstrated remarkable success in modeling human language. Given the similarity of human language and biological sequences such as protein and DNA, researchers have adapted these frameworks to design proteins (21) and, more recently, to generate genomic sequences that contain potentially functional regulatory elements and genes (22–24). Despite these advances, it remains an open question whether language models can be leveraged to efficiently design and analyze complex engineered DNA.

Here, we introduce PlasmidGPT, a generative framework for designing and annotating plasmid DNA sequences (Fig. 1A). Our framework is built on a decoder-only transformer model that is pretrained on 153,208 plasmid sequences from Addgene (25), a public repository for engineered DNA sequences. We demonstrate that sequence embeddings generated by PlasmidGPT encode plasmid sequences into a continuous numerical space. These sequence representations facilitate the visualization of research topics across laboratories by capturing sequence-level similarities and variations. Leveraging simple machine learning models trained on these embeddings, PlasmidGPT enables the fast identification of a wide range of high-level plasmid features (vector type, selectable marker, growth strain, and lab of origin) directly from sequence, facilitating plasmid analysis tasks such as functional annotation and provenance tracking. Moreover, PlasmidGPT generates plasmids that have genetic part distributions similar to those of the training sequences. Conditional plasmid generation can be achieved either by providing a user-specified starting sequence or by fine-tuning the model using special tokens that represent specific vector types. Furthermore, we experimentally validated the functionality of two model-generated plasmids in bacterial cells.

Soluble signals released by neighboring cells direct how the human kidney is built

A human kidney is built once, before birth, and the rules guiding its construction have been largely inferred from animal models. Researchers at Penn Medicine have now read those rules directly in human tissue. They found that developing cells are steered by soluble signals released by their neighbors—and identified insulin-like growth factor 2 (IGF2) as a key signal that sustains the kidney’s stem-like cells.

Their study is published in Nature Genetics and was conducted by the Penn-CHOP Kidney Innovation Center, a research collaboration between the Perelman School of Medicine and Children’s Hospital of Philadelphia, working with bioengineers from Penn’s School of Engineering and Applied Science and the Institute for Regenerative Medicine.

How well a kidney is built before birth matters for the rest of a person’s life. Healthy people are born with anywhere from 200,000 to 2 million nephrons—the microscopic filtering units that clear waste from the blood—and no more are made after birth. A low nephron endowment carries a higher lifetime risk of high blood pressure and chronic kidney disease, and severe disruptions of kidney formation are the leading cause of chronic kidney disease in children.

The Great Progression: Peter Leyden on AI and the Next 25 Years

80 years, America tears itself down and rebuilds. 1776. 1865. 1945. And now, according to Peter Leyden, 2025.

A couple of months ago I sat down with Peter for SingularityFM. He’s the guy who co-wrote the 1997 “Long Boom” cover story for WIRED, and he’s spent three decades trying to map where this civilization is actually headed. His new book, The Great Progression (HarperCollins), makes an audacious claim:

AI, clean energy, and bioengineering are converging into a change on the scale of the Enlightenment. And Trump, in his framing, isn’t the future. He’s the wrecking ball that clears the ground for whatever comes next.

I didn’t let him off easy. I pushed him on the human cost every one of these reinventions has demanded in the past. I pushed him on the myth of golden ages, using Ada Palmer’s work on the Renaissance. I pushed him on Denmark and Norway ripping devices back out of classrooms after being early adopters. He had an answer for all three. Whether they hold up is for you to decide.

What we get into:

The email from Kevin Kelly that pulled Peter into WIRED’s founding years What the Long Boom thesis got right, and what it missed Why he calls Trump a wrecking ball, not a builder The coming “abundance politics” coalition The strange pessimism about AI unique to English-speaking countries How writing this book with AI made him 2 to 3x more productive.

Continue reading “The Great Progression: Peter Leyden on AI and the Next 25 Years” | >

Gene Therapy Improves Stem Cell Transplant Success

Stem cell transplantation (or bone marrow transplantation) and gene therapy have revolutionized the way oncologists treat patients. Both approaches have the potential to cure patients with sickle cell disease, b-thalassemia, immune disorders, and even several blood cancers. There are specific transplants that can occur: autologous and allogenic. Both forms of transplantation are dependent on the source of the donor cells. Autologous transplants use the patient’s healthy stem cells taken from the patient before systemic treatment, like chemotherapy. Allogenic transplants use stem cells collected from healthy donors. Typically stems cells for an allogenic transplant is from a family member or umbilical cord blood to genetically match the recipient. If stem cells are not genetically matched, then the donor will recognize the stem cells as foreign and reject the transplant. This rejection is known as Graft-versus-Host Disease (GvHD). Infusion of stem cells only takes about 15 minutes, but the engraftment or the ability for cells to proliferate and establish themselves in the bone marrow takes up to 18 days. For full immune system recovery, most physicians see patients recover within a year. However, stem cell transplants are not perfect. There is a high rate of GvHD, infection, and disease relapse. Patients also usually receive chemotherapy before a transplant, which increases toxicity and reduces success of the procedure. Scientists are currently working on how to improve the success rate of transplantation.

A recent article in Nature, by Dr. Pietro Genovese and others, demonstrated that stem cell transplantations are safer when chemotherapy is replaced with a targeted treatment. Researchers used antibodies that recognize and target markers on blood-forming stem cells. This approach helps clear harmful preexisting stem cells from the patient before transplantation, instead of using toxic chemotherapeutic agents that damage DNA throughout the body. The antibody approach is more specific and reduces toxicity in patients. Genovese is an Assistant Professor at Dana Farber and Boston Children’s Cancer and Blood Disorders Center. His work focuses on gene-editing and bioengineering technologies that improve stem and immune cells. Specifically, Genovese investigates ways to improve gene therapy and develop approaches to enhance treatment for patients with hematological malignancies.

Previously, an antibody could not discern between current recipient and infused autologous stem cells. Antibodies could also attack transplanted cells from a separate donor (allogenic). To avoid these issues, Genovese and his team have used gene-editing tools to select for a specific biomarker on the surface of donor stem cells. The small edit prevented the antibody from binding to the donor stem cells. These antibodies were able to avoid the healthy stem cells and still eliminate the infected stem cells. This approach allows the donor stem cells to properly graft in the host and improve success of transplantation.

Tailored FcγR blockade enhances immune checkpoint therapy and overcomes resistance

Fc-gamma receptors (FcγRs) regulate IgG antibody activity, and Fc-engineering is a proven method to improve the efficacy of tumor-targeting antibodies. Here, we explore tailored FcγR blockade to enhance the therapeutic efficacy and tolerability of immune checkpoint-blocking (ICB) antibodies.

Mechanistically matched murine surrogate and human lead FcγR-blocking and immune checkpoint-blocking antibodies were used to study whether tailored FcγR-blockade, targeting FcγRIIB selectively or all FcγRs, can enhance the efficacy and overcome resistance to immune checkpoint therapy in vivo and in vitro. Mechanistic studies were performed with clinical reagents, including ipilimumab, nivolumab, pembrolizumab, and human FcγRIIB-selective (BI-1607) and pan-FcγR-blocking (BI-1206) antibodies, using human cells and transgenic animals with clinically relevant expression of immune checkpoint receptors.

We demonstrate that FcγRIIB-selective and pan-FcγR-blocking antibodies increase the in vivo efficacy of αCTLA-4 and αPD-1 antibodies, respectively. FcγRIIB-selective antibody enhancement of αCTLA-4 was associated with increased intratumoral Treg depletion, myeloid reprogramming, interferon-γ and CXCL10-induction, and increased activated effector CD8+ T cells, correlating with higher activating-to-inhibitory (A: I) FcγR engagement ratios. Conversely, pan-FcγR blockade protected αPD-1-coated T cells from macrophage phagocytosis, increasing intratumoral activated CD8+ T cells by decreasing activating and inhibitory FcγRs.

AI‑designed gene‑editing enzymes expand the CRISPR toolbox

Scientists have made many advances using traditional CRISPR technology, especially in medicine, but they are now seeking ways to create genuinely new gene-editing enzymes with properties that have not already evolved naturally. A new study, published in Science, describes a new AI-designed synthetic TnpB enzyme, called SynTnpBs, that has outperformed the natural reference enzyme.

Creating new gene editors CRISPR tools use a programmable guide RNA to direct an enzyme to a specific target in the genome to edit (cut, insert or correct) DNA. The TnpB enzyme is a compact ancestor of certain CRISPR enzymes, called CRISPR-Cas12 enzymes. Researchers think its small size could make it useful in situations where delivery space is limited, like some kinds of gene editing in plants. However, these enzymes can be difficult to redesign.

While AI has been useful for automating complex genome editing and predicting DNA repair outcomes, most AI methods used for generating gene-editing enzymes have produced versions that are still very similar to natural proteins. When researchers have attempted to create new editors with novel, useful properties, they have found it challenging to change the protein without breaking the molecular contacts needed for DNA editing.

Gene editing tool reduces Huntington’s toxic protein fragments and symptoms in mice

A gene-editing tool designed to precisely rewrite the gene that causes Huntington’s disease reduced toxic protein fragments and symptoms associated with the disease in mice, researchers at the University of Illinois Urbana-Champaign report.

While other gene-based treatments have focused on turning the gene off, the Illinois team took a different approach. The researchers designed a base-editing tool to alter a specific point in the huntingtin gene so the cell’s machinery would skip over a small section prone to generating toxic fragments while preserving enough huntingtin protein to support its normal functions.

Led by Pablo Perez-Pinera and Thomas Gaj, professors of bioengineering at the U. of I., the researchers published their findings in the journal Nature Biomedical Engineering.

Insect-inspired electronic nose: Turning semiconductor chips into olfactory sensors

Bioengineers at the University of California San Diego integrated the olfactory receptor of an insect called a jumping bristletail into semiconductor chips made of graphene, creating an electronic nose capable of sniffing out a wide variety of small organic compounds. This biomimetic bioelectronic sensor can detect and distinguish between molecules that are difficult for conventional electronic sensors to differentiate. This work opens the door to building semiconductor-based chemical sensing systems inspired by nature for applications in health care, environmental monitoring, food quality, agriculture and biodefense.

In a paper published in Advanced Materials, researchers led by bioengineers at UC San Diego describe a method for manufacturing the MhOR5 odorant receptor from the insect Machilis hrabei at scale and chemically attaching the purified MhOR5 protein to high-performance graphene field effect transistors (gFETs). gFETs are semiconductor devices that rely on graphene instead of silicon as the conductive material, resulting in exceptional sensitivity to molecular changes.

The researchers tested their MhOR5-functionalized gFETs against 16 chemically diverse compounds, including DEET, hexanol and eugenol, at different concentrations. The sensor produced a concentration-dependent electrical response for each of the 16 compounds.

AI finds tiny gene editor changes that reduce unintended DNA edits

Gene editing is a highly precise and powerful technology that allows scientists to insert, delete, modify or replace DNA bases in living organisms. It has a variety of uses, including correcting disease-causing mutations and improving crops. Tools like CRISPR act as molecular scissors that target specific places in a genome to make these changes. But the technology is not perfect and can accidentally edit the wrong pieces of DNA or RNA.

In research published in Nature, scientists describe a new framework that uses AI to make these tools more accurate. Hoi Yee Chu and Alan S.L. Wong of the University of Hong Kong published a News and Views piece in the same journal on the significance of this research.

Inherited gene variants may shape CAR-T therapy benefits and toxic side effects

Chimeric antigen receptor (CAR)-T cell therapy, which reprograms an individual’s immune cells to seek out and destroy certain cancer cells, has revolutionized treatment for blood cancers such as lymphoma. But in some patients, the treatment can cause serious side effects. New research led by investigators at the Mass General Brigham Cancer Institute, the Broad Institute of MIT and Harvard, and Dana-Farber Cancer Institute has shown that patients’ inherited genetic makeup can influence whether they benefit from CAR-T cell therapy or experience toxicity from the treatment. The results are published in Science Immunology.

“These findings have important implications for understanding how CAR-T cells behave in patients since each CAR-T cell product is unique to the person from whom it is manufactured, unlike all prior forms of therapy, which are identical across patients,” said lead author Mark B. Leick, M.D., an oncologist at the Mass General Brigham Cancer Institute.

For the study, Leick and his colleagues sequenced the entire genomes of more than 200 patients with aggressive lymphoma from two major clinical trials of CAR-T cell therapy. In one of the trials, patients with T cells with variants that silenced the STXBP2 gene tended to experience toxicity related to CAR-T cell therapy. Also, donor T cells engineered to lack STXBP2 and/or express these STXBP2-silencing variants triggered inflammation.