<div id="readability-page-1"><ul><li>Legal Issues Behind It, IBIO Stock May Be a Long Play For Risk-Tolerant Investors  </li></ul><p>iBio <span>NASDAQ: IBIO<span></span></span> is seeking to narrow what it views as a growing “trust gap” between the promise of generative AI in drug discovery and the high cost and long timelines of bringing a drug to market, according to Alex Taguchi, the company’s Director of Machine Learning. In a webinar presentation on “Generative Antigen Design for GPCR Antibody Discovery,” Taguchi outlined how iBio is integrating generative AI with experimental validation to engineer antigens designed to steer antibody discovery toward specific epitopes and difficult target classes such as G protein-coupled receptors (GPCRs).</p><h2>Integrating AI with structural and functional validation</h2> <p>Taguchi framed the talk around practical questions he said teams must answer before relying on AI in drug development: whether AI is the right tool for a given problem, which models to use, and how to trust model outputs. To address those questions, he said iBio has spent roughly the past 1.5 to two years “deeply integrating” its AI platform into an “integrated experimental structural and functional validation pipeline” to pre-validate AI-generated molecules before deploying them in drug discovery campaigns. </p><p>Taguchi described iBio as an antibody drug discovery company focused on obesity, emphasizing speed and molecule quality. He referenced the company’s IBIO-600 campaign, which he said went “from antibody discovery to development candidate in as little as seven months,” and noted that iBio announced that IBIO-600 was “now approved to go into clinic,” describing a roughly two-year path “from inception all the way to clinic approval.” He also highlighted IBIO-610, a “potential first-in-class Activin E antibody” that iBio is developing for obesity-related indications.  </p><h2>“Personalized baits”: engineered epitopes as discovery tools</h2> <p>To explain iBio’s approach, Taguchi compared traditional antibody discovery to a “fishing expedition,” with large pharmaceutical companies relying on brute force scale. He contrasted that with de novo antibody design, which he said can be akin to “building a lab-grown version of a fish.” iBio’s approach, he said, is to use machine learning to build “personalized baits”—engineered antigens that present specific epitopes in native-like conformations—in order to increase the efficiency of discovering epitope-selective antibodies from human libraries. </p><p>Taguchi said the company uses generative AI to build peptide scaffolds that support linear or discontinuous epitopes, creating small structural mimics that can be used as “bait” in selection workflows. He added that iBio’s antigen designs are optimized for antibody drug discovery, describing them as water-soluble, structurally stable, built with minimal scaffolds to reduce off-target binding, and with scaffold composition “predicted to have low immunogenicity.”  </p><h2>Breaking immune tolerance: latent TGF-beta 1 and Activin E</h2> <p>As a case study for epitope-targeted immunization, Taguchi discussed work with a collaborator pursuing antibodies against latent TGF-beta 1, an immune-modulating target in oncology. He said mouse immunizations had failed because human and mouse latent TGF-beta 1 share 89% sequence identity, making immune tolerance difficult to overcome. iBio produced multiple engineered epitope designs and then validated them experimentally, including binding to a benchmark antibody (SRK-181) and binding to integrin for an integrin-site epitope. </p><p>Taguchi also described an immunization strategy intended to improve translation from peptide binders to full-length protein binders: alternating between peptide immunizations and full-length protein immunizations. He said the campaign produced “good serum titers across the board,” and he highlighted what he called an “eye-opening” result: in some cases, immunizing with engineered epitopes alone produced immune responses similar to co-immunization with full-length protein, suggesting the mimics could be structurally accurate enough to reduce dependence on full-length antigen. </p><p>He then described Activin E as a more challenging obesity-related target, with ~97% sequence identity between human and mouse and difficulty producing active protein due to many disulfide bonds. iBio divided Activin E into six epitope regions and used a ferritin nanoparticle display system, which Taguchi said can present “up to about 50 epitopes per nanoparticle,” to improve immune activation. He reported that immunizing with certain engineered epitopes generated significant immune responses against Activin E, aligning with computational predictions of immunogenic “hotspots.” Taguchi said iBio also observed that immune-tolerance breaking translated to VHH discovery in llamas, including evidence of boosted titers when engineered-epitope immunization was followed by Activin E immunization. </p><p>Taguchi said a key learning was the apparent structural accuracy and improved “zero-shot” performance of diffusion models relative to earlier protein design eras. He also said the small size of engineered epitopes can be advantageous for nanoparticle immunization by enabling higher epitope density.  </p><h2>Extending the platform to GPCRs and junctional targets</h2> <p>Turning to GPCRs, Taguchi described early work targeting CCR8 by designing scaffolds that support extracellular loops; he said experimental NMR structure alignment “aligned really well” with the generative design. He then discussed efforts to create soluble GPCR surrogates that include full extracellular domains by “reimagining the transmembrane domain into a soluble format.” </p><p>Using the GIP receptor as a test case, Taguchi said iBio used a cryo-EM structure as input and generated multiple soluble designs, reporting that the engineered receptor bound its native ligand GIP, did not bind GLP-1, and also bound peptide therapeutics including tirzepatide (Zepbound) and retatrutide. He added that negative stain electron microscopy provided low-resolution support for the expected morphology, though he noted flexibility at the fusion hinge used in imaging. </p><p>As a more difficult challenge, Taguchi described work on GPR75, where he said no experimental structures were available at the time and the team relied on an AlphaFold model as the design input. He said iBio built 12 designs and observed CCL5 binding for one design, along with binding to a commercial anti-GPR75 antibody, and he described negative stain images showing a morphology consistent with expectations.  </p><h2>Amylin receptor case study and selection for cross-reactivity</h2> <p>Taguchi closed with a drug discovery example focused on amylin receptor agonism. He described the amylin receptor as a heterodimeric complex (calcitonin receptor plus RAMP accessory proteins) and said existing peptide agonists can lack “exquisite selectivity” for the amylin receptor pathway. iBio’s goal, he said, was to develop an antibody selective for the receptor complex, then tether a synthetic amylin peptide agonist to produce a selective agonist profile. </p><p>Taguchi said iBio validated engineered antigens for binding behavior consistent with expectations and pursued both immunizations and in vitro selections. In phage panning, he reported that panning against engineered antigens produced “tens of antibody hits” that translated to cell binders, while panning on cells alone yielded no specific binders. </p><p>He also highlighted an in vitro strategy proposed by colleague Cody involving multi-dimensional mammalian display sorting to “bake in” species and subtype cross-reactivity early. Taguchi said the team selected rare clones showing binding profiles across human amylin receptor subtypes and rat cross-reactivity while remaining negative for human calcitonin receptor binding. In an agonism assay after tethering a peptide, he said the construct showed activation of amylin receptor cells without activation of calcitonin receptor cells, contrasting that with a nonspecific profile he attributed to cagrilintide.  </p><h2>Machine learning outlook: engineering epitope and paratope together</h2> <p>In a Q&amp;A exchange, Taguchi said he expects a key hurdle over the next three to five years to be combining de novo antibody design with antigen engineering, arguing that the field is “way too focused” on designing antibodies alone. He said simultaneous engineering of the epitope and paratope could unlock the next wave of innovation. </p><p>Asked about generative model failure modes, Taguchi said a common issue is insufficiently constraining diffusion models to create compact, globular structures. Without high-level structural constraints, he said designs can “hallucinate” non-compact forms that “don’t look like they’ll fold well,” adding that models “still need to be babysat.”</p><h2>About iBio <span>NASDAQ: IBIO<span></span></span></h2><p>iBio, Inc, a biotechnology company, provides contract development and manufacturing services to collaborators and third-party customers in the United States. The company operates in two segments: Biopharmaceuticals and Bioprocessing. Its lead therapeutic candidate is IBIO-100 that is being advanced for investigational new drug development for the treatment of systemic scleroderma and idiopathic pulmonary fibrosis. The company is also developing vaccine candidates comprising IBIO-200 and IBIO-201, which are in preclinical development for the prevention of severe acute respiratory syndrome coronavirus 2; and IBIO-400 for the treatment of classical swine fever.</p><h2>See Also</h2><ul><li>Five stocks we like better than iBio</li></ul><p><em>This instant news alert was generated by narrative science technology and financial data from MarketBeat in order to provide readers with the fastest reporting and unbiased coverage. Please send any questions or comments about this story to contact@marketbeat.com.</em></p><h2>Should You Invest $1,000 in iBio Right Now?</h2><p>Before you consider iBio, you'll want to hear this.</p><p>MarketBeat keeps track of Wall Street's top-rated and best performing research analysts and the stocks they recommend to their clients on a daily basis. MarketBeat has identified the five stocks that top analysts are quietly whispering to their clients to buy now before the broader market catches on... and iBio wasn't on the list.</p><p>While iBio currently has a Buy rating among analysts, top-rated analysts believe these five stocks are better buys.</p><p>View The Five Stocks Here</p>                  </div>

LABU

SBIO

IBIO

IBBQ

<p>iBio (NASDAQ:IBIO) is addressing the "trust gap" in drug discovery by integrating generative AI with experimental validation to enhance GPCR antibody discovery. Director Alex Taguchi highlighted the company's approach in a webinar, emphasizing the development of engineered antigens to improve antibody discovery efficiency. iBio's recent campaigns, including IBIO-600 and IBIO-610, showcase rapid progression from discovery to clinical approval. The company aims to optimize antibody drug discovery through personalized baits and innovative immunization strategies, demonstrating promising results in overcoming immune tolerance and enhancing immune responses against challenging targets like Activin E.</p>

Market Beat

Invesco Biotechnology & Genome ETF

Direxion S&P Biotech Bull 3X

ALPS Medical Breakthrough

First Trust NYSE Arca Biotech Index Fd

iBio

VanEck Biotech ETF

Invesco Nasdaq Biotechnology ETF

iShares Biotechnology ETF

Proshares Ultra Nasdaq Biotech

SPDR S&P Biotech

iBio Details Generative AI Antigen Platform to Speed GPCR Antibody Discovery and Close “Trust Gap”