Optimal tip sonication settings, namely tip position, input power, and pulse durations, are necessary for temperature sensitive procedures like preparation of viable cell extract. In this paper, the optimum tip immersion depth (20-30% height below the liquid surface) is estimated which ensures maximum mixing thereby enhancing thermal dissipation of local cavitation hotspots. A finite element (FE) heat transfer model is presented, validated experimentally with (R2 > 97%) and used to observe the effect of temperature rise on cell extract performance of E. coli BL21 DE3 star strain and estimate the temperature threshold. Relative yields in the top 10% are observed for solution temperatures maintained below 32°C; this reduces below 50% relative yield at temperatures above 47°C. A generalized workflow for direct simulation using the COMSOL code as well as master plots for estimation of sonication parameters (power input and pulse settings) is also presented.

Optimal tip sonication settings, namely tip position, input power, and pulse durations, are necessary for temperature sensitive procedures like preparation of viable cell extract. In this paper, the optimum tip immersion depth (20–30% height below the liquid surface) is estimated which ensures maximum mixing thereby enhancing thermal dissipation of local cavitation hotspots. A finite element (FE) heat transfer model is presented, validated experimentally with (R2 > 97%) and used to observe the effect of temperature rise on cell extract performance of Escherichia coli BL21 DE3 star strain and estimate the temperature threshold. Relative yields in the top 10% are observed for solution temperatures maintained below 32°C; this reduces below 50% relative yield at temperatures above 47°C. A generalized workflow for direct simulation using the CONSOL code as well as master plots for estimation of sonication parameters (power input and pulse settings) is also presented.

Reinforcement learning (RL), a subset of machine learning (ML), can potentially optimize and control biomanufacturing processes, such as improved production of therapeutic cells. Here, the process of CAR-T cell activation by antigen presenting beads and their subsequent expansion is formulated in-silico. The simulation is used as an environment to train RL-agents to dynamically control the number of beads in culture with the objective of maximizing the population of robust effector cells at the end of the culture. We make periodic decisions of incremental bead addition or complete removal. The simulation is designed to operate in OpenAI Gym which enables testing of different environments, cell types, agent algorithms and state-inputs to the RL-agent. Agent training is demonstrated with three different algorithms (PPO, A2C and DQN) each sampling three different state input types (tabular, image, mixed); PPO-tabular performs best for this simulation environment. Using this approach, training of the RL-agent on different cell types is demonstrated, resulting in unique control strategies for each type. Sensitivity to input noise (sensor performance), number of control step interventions, and advantage of pre-trained agents are also evaluated. Therefore, we present a general computational framework to maximize the population of robust effector cells in CAR-T cell therapy production.

No abstract

Optimal tip sonication settings, namely tip position, input power, and pulse durations, are necessary to ensure proper mixing and maintain solution temperature below a critical temperature. This is significant for temperature sensitive procedures like preparation of viable cell extract for in vitro protein synthesis. In this paper, the optimum tip immersion depth is estimated which ensures maximum mixing thereby enhancing thermal dissipation of local cavitation hotspots inside the sonication tube; from modeled velocity field streamlines this is found at immersion depths between 20-30% height below the liquid surface. A simplified finite element (FE) heat transfer model is presented and validated experimentally with (R2 > 97%) which can predict the temperature rise over time in a tip-sonicated vessel. This model is used to observe the effect of temperature rise on cell extract performance of E. coli BL21 DE3 star strain and estimate the temperature threshold. From the combined heat map of yield and temperature it is observed that yield is correlated with final steady state temperature. Relative yields in the top 10% are observed for solution temperatures maintained below 32°C; this reduces below 50% relative yield at temperatures above 47°C. To extend utility of these finite element models to other temperature sensitive sonication processes, we also present a generalized workflow for direct simulation using the FE code as well as master plots for estimation of sonication parameters (power input and pulse settings) without need of running the code.