The Defense Threat Reduction Agency (DTRA) is seeking innovative solutions through the Small Business Innovation Research (SBIR) program to develop machine learning models for predictive engineering and optimization in biomining essential resources. The objective is to create a computational tool that leverages advances in synthetic biology and AI/ML to enhance the selective extraction and recovery of critical materials, such as structural metals and rare earth elements, which are vital for high-tech and clean-energy sectors. This initiative is particularly important as it aims to provide environmentally friendly alternatives to traditional mining methods while addressing the growing threats posed by advancements in synthetic biology and materials science in the Weapons of Mass Destruction (WMD) spectrum. Interested parties should note that the solicitation is currently closed, with the application due date having been October 23, 2024, and further details can be found at the DOD SBIR/STTR website.

## Description

The developed software tool will augment the suite of tools available to the Defense Threat Reduction Agency (DTRA) and end users for environmental monitoring and sampling, and increase understanding of processes required to degrade niche materials. Pacing threat countries continue to integrate techniques in the fields of synthetic biology, materials science, and AI/ML for developing advanced technologies in the Weapons of Mass Destruction (WMD) threat spectrum. DTRA seeks disruptive approaches in biomining (also referred to as bioleaching) for using microorganisms to extract or degrade metals, minerals, and other materials.Approaches in biomining have been sought as an alternative to traditional mining and are generally considered to be more environmentally friendly. This topic aims to leverage advances in synthetic biology to develop a computational tool to explore the parameter spaces of engineered microbes, and to predict desired material selectivity and degradation performance. Outcomes of the effort will also drive improvements in the biomining field for improving the efficiency of processing times and material recovery yields. Target materials of interest include structural metals such as iron, copper, and nickel; rare earth elements (REE) for electronics, in addition to critical minerals and other critical materials (e.g. silicon and manganese) [1-4]. Automated, large-scale exploration of biological designs may be realized through a combination of explorative and exploitative algorithmic means. Advancements in AI/ML have revolutionized the way researchers manipulate biological systems. AI/ML models can predict and optimize metabolic pathways with unprecedented precision, and streamline the identification of key genes, enzymes, and metabolites, facilitating the design of more efficient and tailored biological pathways to produce valuable compounds and materials [5-7]. Such products include biofuels, pharmaceuticals, and specialty chemicals. For instance, researchers from the Massachusetts Institute of Technology (MIT) introduced &ldquo;BioAutoMATED,&rdquo; an automated ML platform for designing biological sequences and includes an interpretation module for thorough model analysis [8]. Another research group at the University of Texas at Austin (UT Austin) addressed the ecological challenge of solid plastic waste, and the potential green and scalable option for waste recycling of poly(ethylene terephthalate) (PET) [9]. The UT Austin team developed and demonstrated a structure-based ML platform to engineer a robust and active PET hydrolase [9].

## Objective

Produce a viable framework for engineering designs using approaches in synthetic biology, biomining, and AI/ML that guide selective extraction and recovery of essential resources used in high-tech and clean-energy sectors. This topic seeks development of (1) computational methods for engineering the design of microbial consortia and/or enzymes with optimal degradation activity for target materials; (2) realize a scalable software system with automated capabilities that may inform in vitro experimental designs.

## Phase I

Leverage computational methods (e.g. AI/ML, in silico models) to inform and guide potential biomining agents that may be employed in vitro. The software architecture shall be scalable, and ideally, shall leverage automated annotation capabilities in lieu of human annotation. Performers are encouraged to use their own data, if available, including &ldquo;wet lab&rdquo; experimental data for training and testing ML models. The software shall enable a user to incorporate and process additional data on an ad hoc basis.Phase I deliverables will include (1) a final report and (2) demonstration of the preliminary architecture.&nbsp; The report should provide results on architecture performance using unambiguous statistical methods, describe training and development, and identify advantages, limitations, and weaknesses.&nbsp; The architecture shall be described, including operating system, other software requirements (if applicable), and data sources.

## Phase II

Phase II efforts will focus on refinement of the approach developed during Phase I and prototype demonstration.&nbsp; The Phase II deliverables will be a prototype demonstration of the neural network architecture and a report detailing: (1) a description of the approach, optimization techniques, and performance outcomes; (2) training, testing and validation methods; (3) a real world evaluation of the approach with a use case of mutual interest with DTRA; and (4) advantages, disadvantages, and limitations of the approach.&nbsp; The performer will identify weaknesses of the approach, and identify methods that may improve performance in the classifier and aspects of the overall architecture.&nbsp; The performer will provide details about user interfaces (if applicable) and any associated executables.

## Phase III: Dual Use Applications

The performer will identify and employ features that have the potential for use in commercial applications.

Topic

Agency

Program

Topic Link

Additional Information