GUIDE’s integrated computational platform

Currently, the majority of early GUIDE tool development is focused on monoclonal antibodies due to their broad applicability across threat space. Over the course of the program, GUIDE will progressively broaden the types of MCMs that can be computationally discovered and designed. This will involve an increased emphasis on vaccines computationally designed to induce broadly protective immune responses, and small molecule MCMs that may have applicability to both biological and chemical threats. In addition, there will be computational approaches that bridge discovery with advanced manufacturing processes and with models to expedite pre-clinical testing.

 

Our approach

Revolutionary acceleration of drug development requires a shift from traditional, lengthy experimental approaches to advanced computational tools. The GUIDE drug discovery approach is focused around three main activities:

  • Co-design, or the process of computationally proposing new drug candidates, while simultaneously optimizing for safety, efficacy, manufacturability and pharmacokinetics/pharmacodynamics.
  • Test, or the process of making the drug candidates in the laboratory and testing their effectiveness.
  • Learn, each time we complete this cycle we learn from the process and have new data to train our machine learning models, enabling continuous improvement of our drug design process.

An innovative approach to drug discovery

Computational System

 

GUIDE is currently in standup phase, as we’re working to integrate 21 tools into an advanced computational system. This process will enable co-optimization decision systems that manage a variety of property prediction models.

Many of the planned tools, particularly those based on machine learning, are dependent on currently non-existent training data. The GUIDE program is producing first-in-kind data sets to build these models. Advanced computational approaches being developed for GUIDE and substantial laboratory infrastructure to support high-throughput data generation for model building, MCM analysis and characterization, and rapid MCM response. 

Streamlining drug development

GUIDE’s longer-term goal is to develop computational approaches that will not rely on existing ‘near-neighbor’ MCMs for retargeting. A fully de novo discovery and design capability will start with pathogen sequence information, correctly predict structurally correct targets, and design highly potent MCMs based on first principles, simulations, and computational models. This is a revolutionary paradigm change in MCM development not just for the biothreat space, but for the pharmaceutical industry at large.

GUIDE optimizes medical countermeasures by focusing on safety (including immunogenicity, immunotoxicity, specificity, effector function, and TCR), efficacy (including binding/targets, effector function, dosing regimen, immunomodulation, and potency), manufacturing (including aggregation, glycosylation, solubility, formulation, and stability), and pharmacokinetics/pharmacodynamics (including delivery, bioavailability, route of administration, and half-life).
Critical quality attributes streamline drug development by focusing on a nexus of safety, efficacy, manufacturing, and pharmacokinetics/pharmacodynamics.

 

GUIDE is developing approaches to optimize MCMs across a complex set of product critical quality attributes including safety, efficacy, manufacturability, and pharmacokinetics/pharmacodynamics (PK/PD). Validated computational models and approaches can significantly reduce trial-and-error experimentation and greatly increase the likelihood that the MCM will meet the desired target product profile in days to weeks rather than decades.

Preempting threats with prototype preparedness

GUIDE’s interagency approach allows for computational ‘retargeting’ of existing MCMs. Prototypes represent larger threat families, or groups of toxins, bacteria, or viruses. If the biothreat space is a vast ocean and each threat is an island, threat families can be thought of as archipelagos of related threats.

Deep knowledge about one or more prototype islands in the archipelago and availability of MCMs to address those threats allow GUIDE to preemptively retarget MCMs across the archipelago. This method can be used to rapidly respond to an unanticipated threat such as an engineered virus related to an existing viral prototype.

 

Prototype Archipelago

 

Prototype preparedness requires a comprehensive view of the threat landscape, an expert-informed approach to prototype selection, and the continued refinement of retargeting tools and models to create and validate MCM solutions. The combination of strategically positioned prototype islands with advanced high-performance computing for biologics and small molecule development allows GUIDE to match the dynamic nature of the evolving threat landscape and offers the potential for fully preemptive preparedness across a range of threats.

Join the GUIDE team

From immunology to machine learning and systems biology to molecular dynamics simulation, GUIDE is a multidisciplinary effort. Discover open positions at Lawrence Livermore National Laboratory.