Explore the vast range of titles available.

We present a protein engineering approach to directed evolution with machine learning that integrates a new semi-supervised neural network fitness prediction model, Seq2Fitness, and an innovative optimization algorithm, b iphasic a nnealing for d iverse and a daptive s equence s ampling (BADASS) to design sequences. Seq2Fitness leverages protein language models to predict fitness landscapes, combining evolutionary data with experimental labels, while BADASS efficiently explores these landscapes by dynamically adjusting temperature and mutation energies to prevent premature convergence and to generate diverse high-fitness sequences. Compared to alternative models, Seq2Fitness improves Spearman correlation with experimental fitness measurements, increasing from 0.34 to 0.55 on sequences containing mutations at positions entirely not seen during training. BADASS requires less memory and computation compared to gradient-based Markov Chain Monte Carlo methods, while generating more high-fitness and diverse sequences across two protein families. For both families, 100% of the top 10,000 sequences identified by BADASS exceed the wildtype in predicted fitness, whereas competing methods range from 3% to 99%, often producing far fewer than 10,000 sequences. BADASS also finds higher-fitness sequences at every cutoff (top 1, 100, and 10,000). Additionally, we provide a theoretical framework explaining BADASS’s underlying mechanism and behavior. While we focus on amino acid sequences, BADASS may generalize to other sequence spaces, such as DNA and RNA.

Thermal energy management in metal-organic frameworks (MOFs) is an important, yet often neglected, challenge for many adsorption-based applications such as gas storage and separations. Despite its importance, there is insufficient understanding of the structure-property relationships governing thermal transport in MOFs. To provide a data-driven perspective into these relationships, here we perform large-scale computational screening of thermal conductivity k in MOFs, leveraging classical molecular dynamics simulations and 10,194 hypothetical MOFs created using the ToBaCCo 3.0 code. We found that high thermal conductivity in MOFs is favored by high densities (> 1.0 g cm−3), small pores (< 10 Å), and four-connected metal nodes. We also found that 36 MOFs exhibit ultra-low thermal conductivity (< 0.02 W m−1 K−1), which is primarily due to having extremely large pores (~65 Å). Furthermore, we discovered six hypothetical MOFs with very high thermal conductivity (> 10 W m−1 K−1), the structures of which we describe in additional detail.

Tuberculosis (TB) remains one of the main causes of human death around the globe. The mortality rate for patients infected with active TB goes beyond 50% when not diagnosed. Rapid and accurate diagnostics coupled with further prompt treatment of the disease is the cornerstone for controlling TB outbreaks. To reduce this burden, the existing gap between detection and treatment must be addressed, and dedicated diagnostic tools such as biosensors should be developed. A biosensor is a sensing micro-device that consists of a biological sensing element and a transducer part to produce signals in proportion to quantitative information about the binding event. The micro-biosensor cell considered in this investigation is designed to operate based on aptamers as recognition elements against Mycobacterium tuberculosis secreted protein MPT64, combined in a microfluidic-chamber with inlet and outlet connections. The microfluidic cell is a miniaturized platform with valuable advantages such as low cost of analysis with low reagent consumption, reduced sample volume, and shortened processing time with enhanced analytical capability. The main purpose of this study is to assess the flooding characteristics of the encapsulated microfluidic cell of an existing micro-biosensor using Computational Fluid Dynamics (CFD) techniques. The main challenge in the design of the microfluidic cell lies in the extraction of entrained air bubbles, which may remain after the filling process is completed, dramatically affecting the performance of the sensing element. In this work, a CFD model was developed on the platform ANSYS-CFX using the finite volume method to discretize the domain and solving the Navier–Stokes equations for both air and water in a Eulerian framework. Second-order space discretization scheme and second-order Euler Backward time discretization were used in the numerical treatment of the equations. For a given inlet–outlet diameter and dimensions of an in-house built cell chamber, different inlet liquid flow rates were explored to determine an appropriate flow condition to guarantee an effective venting of the air while filling the chamber. The numerical model depicted free surface waves as promoters of air entrainment that ultimately may explain the significant amount of air content in the chamber observed in preliminary tests after the filling process is completed. Results demonstrated that for the present design, against the intuition, the chamber must be filled with liquid at a modest flow rate to minimize free surface waviness during the flooding stage of the chamber.

Metal-organic frameworks (MOFs) are emerging as a promising class of materials for applications such as gas storage, separation, and catalysis, attributed to their large surface area, tunable pore geometry, and high porosity. However, their thermal transport properties have been relatively underexplored, leaving a gap in our understanding of the relationship between structure and thermal conductivity - knowledge that is crucial for the design of MOFs with specific thermal transport properties. To bridge this gap, we performed the first computational high-throughput screening of over 10,000 hypothetical MOFs using classical molecular dynamics simulations and the Green-Kubo method. Our research also includes an investigation of the impact of both randomly and symmetrically distributed defects on the thermal conductivity of two well-known MOFs, UiO-66 and HKUST-1. The results indicate that while randomly introduced missing linker and missing cluster defects generally reduce thermal conductivity, spatially correlated missing linker defects can actually increase thermal conductivity when carefully incorporated into the parent framework. Given that approximately 90,000 synthesized and 500,000 predicted MOFs are known, there remains a vast, largely unexplored MOF-thermal conductivity structure-property design space, primarily due to the high computational cost of molecular dynamics simulations. To circumvent this challenge, we trained several graph neural network models to rapidly predict the thermal conductivity tensor of MOFs, thus facilitating the exploration of the MOF design space.In conclusion, this dissertation provides critical insights into the design of MOFs with tailored thermal properties and underscores the importance of considering structural features and defects in the design of thermally conductive MOFs for a variety of applications.

We present a novel protein engineering approach to directed evolution with machine learning that integrates a new semi-supervised neural network fitness prediction model, Seq2Fitness, and an innovative optimization algorithm, biphasic annealing for diverse adaptive sequence sampling (BADASS) to design sequences. Seq2Fitness leverages protein language models to predict fitness landscapes, combining evolutionary data with experimental labels, while BADASS efficiently explores these landscapes by dynamically adjusting temperature and mutation energies to prevent premature convergence and find diverse high-fitness sequences. Seq2Fitness predictions improve the Spearman correlation with fitness measurements over alternative model predictions, e.g., from 0.34 to 0.55 for sequences with mutations residues that are absent from the training set. BADASS requires less memory and computation compared to gradient-based Markov Chain Monte Carlo methods, while finding more higher-fitness sequences and maintaining sequence diversity in protein design tasks for two different protein families with hundreds of amino acids. For example, for both protein families 100% of the top 10,000 sequences found by BADASS have higher Seq2Fitness predictions than the wildtype sequence, versus a broad range between 3% to 99% for competing approaches with often many fewer than 10,000 sequences found. The fitness predictions for the top, top 100th, and top 1,000th sequences found by BADASS are all also higher. In addition, we developed a theoretical framework to explain where BADASS comes from, why it works, and how it behaves. Although we only evaluate BADASS here on amino acid sequences, it may be more broadly useful for exploration of other sequence spaces, including DNA and RNA. To ensure reproducibility and facilitate adoption, our code is publicly available here. Designing proteins with enhanced properties is essential for many applications, from industrial enzymes to therapeutic molecules. However, traditional protein engineering methods often fail to explore the vast sequence space effectively, partly due to the rarity of high-fitness sequences. In this work, we introduce BADASS, an optimization algorithm that samples sequences from a probability distribution with mutation energies and a temperature parameter that are updated dynamically, alternating between cooling and heating phases, to discover high-fitness proteins while maintaining sequence diversity. This stands in contrast to traditional approaches like simulated annealing, which often converge on fewer and lower fitness solutions, and gradient-based Markov Chain Monte Carlo (MCMC), also converging on lower fitness solutions and at a significantly higher computational and memory cost. Our approach requires only forward model evaluations and no gradient computations, enabling the rapid design of high-performing proteins that can be validated in the lab, especially when combined with our Seq2Fitness models. BADASS represents a significant advance in computational protein engineering, opening new possibilities for diverse applications.