Search bioRxiv⌕ Search

Biology subjects

Benson, J. D.

Publications and source records attributed to Benson, J. D..

4 recordsLinked to original sources

Machine learning and hypothesis driven optimization of bull semen cryopreservation media

Cryopreservation provides a critical tool for dairy herd genetics management. Due to widely varying inter- and within-bull post thaw fertility, recent research on cryoprotectant extender medium has not dramatically improved suboptimal post-thaw recovery in industry. This progress is stymied by the interactions between samples and the many components of extender media and is often compounded by industry irrelevant sample sizes. To address these challenges, here we demonstrate blank-slate optimization of bull sperm cryopreservation media by supervised machine learning. We considered two supervised learning models: artificial neural networks and Gaussian process regression (GPR). Eleven media components and initial concentrations were identified from publications in bull semen cryopreservation, and an initial 200 extender-post-thaw motility pairs were used to train and 32 extender-post-thaw motility pairs to test the machine learning algorithms. The median post-thaw motility after coupling differential evolution with GPR the increased from 52.6 {+/-} 6.9% to 68.3 {+/-} 6.0% at generations 7 and 17 respectively, with several media performing dramatically better than control media counterparts. This is the first study in which machine learning was used to determine the best combination of constituents to optimize bull sperm cryopreservation media, and provides a template for optimization in other cell types.

bioengineering↗

Osmotic, temperature, and cytotoxic damage in sea urchin oocytes

Sea urchin (Paracentrotus lividus) oocytes are an important species for aquaculture and as a model species for multiple scientific fields. Despite their importance, methods of cryopreserved biobanking of oocytes are currently not possible. Optimized cryoprotectant loading may enable vitrification methods of cryopreservation and thus long-term storage of oocytes. Determining an optimized protocol requires membrane characteristics and models of damage associated with the vitrification loading protocol, namely osmotic, temperature, and cytotoxic damage. We present and experimentally evaluated state-of-the-art models alongside our novel models. We experimentally verify the damage models throughout time at difference treatment intensities. Osmotic damage experiments consisted of hypertonic solutions composed of seawater supplemented with NaCl or sucrose and hypotonic solutions composed of seawater diluted with deionized water. Treatment times ranged from 2 to 30 minutes. To test temperature damage (in particular chill injury), oocytes were exposed to 1.7 {degrees}C, 10 {degrees}C, and 20 {degrees}C (control) for exposure times ranging from 2 to 90 minutes. Cytotoxicity was investigated by exposing oocytes to solutions of Me2SO for exposure times ranging from 2 to 30 minutes. We identify appropriate models and use these to search for an optimal loading protocol, namely the time dependent osmotic damage model (for osmotic damage), the temperature dependent model (for temperature damage), and the external molality Arrhenius power model (for cytotoxicity). We combined these models to estimate total damage during a cryopreservation loading protocol and performed a exhaustive grid search for optimal loading for a given goal intracellular cryoprotectant concentration. Given our fitted models, we find sea urchin oocytes can only be loaded to 0.13 Me2SO v/v with a 50% survival, For reference, levels for vitrification are approximately 0.45 v/v. Our synthesis of damages is the first of its kind, and enables a fundamentally novel approach to modelling survival for cells in general.

cell biology↗

An agent based model of intracellular ice formation and propagation in small tissues

Successful cryopreservation of tissues and organs would be a critical tool to accelerate drug discovery and facilitate myriad life saving and quality of life improving medical interventions. Unfortunately success in tissue cryopreservation is quite limited, and there have been no reports of successful long term organ cryopreservation. One principal challenge of tissue and organ cryopreservation is the propagation of damaging intracellular ice. Understanding the probability that cells in tissues form ice under a given cryopreservation protocol would greatly accelerate protocol design, enabling rational model-based decisions of all aspects of the cryopreservation procedure. Established models of intracellular ice formation (IIF) in individual cells have previously been extended to small linear (one-cell-wide) arrays to establish the theory of intercellular ice propagation in tissues. However these small-scale lattice-based tissue ice propagation models have not been extended to more realistic tissue structures, and do not account for intercellular forces that arise from the expansion water into ice that may cause mechanical disruption of tissue structures during freezing. To address these shortcomings, here we present the development and validation of a lattice-free agent-based stochastic model of ice formation and propagation in small tissues. We validate our Monte Carlo model against Markov chain models in the linear two-cell and four-cell arrays presented in the literature, as well as against new Markov chain results for 2 x 2 arrays. Moreover we expand the existing model to account for the solidification of water into ice in cells. We then use literature data to inform a model of ice propagation in hepatocyte disks, spheroids, and tissue slabs. Our model aligns well with previously reported experiments, and demonstrates that the mechanical effects of individual cells freezing can be captured. Author summaryThe widespread ability to successfully store, or cryopreserve, tissues and organs in liquid nitrogen temperatures would be game changing for human and animal medicine and drug discovery. However, success is limited to a select number of small tissues, and no organs can currently be stored in a frozen or solid state and survive thawing. One major contributor to damage during this process is the formation of intracellular ice, and its associated cell level damage. This ice formation is complicated in tissues by the number of intercellular connections facilitating intercellular ice propagation. Previous researchers have developed and experimentally validated simple one dimensional models of ice propagation in tissues, but these fail to capture complex tissue geometries, and have many fewer intercellular connections compared to three dimensional tissues. In this paper, we adopt previous models of ice formation and propagation to a model capable of capturing arbitrary cell orientations in three dimensions, allowing for realistic tissue structures to be modelled. We validated this tool on simple models and with experimental data, and then test it on three structures made of digital liver cells: disks, spheroids, and slabs. We show that we can capture new information about the interaction of cooling the tissue, the formation of intracellular ice, the movement of ice from one cell to another, and the mechanical disruption that occurs during this process. This allows for novel insights into a mechanism of damage during cryopreservation that is cooling rate and tissue structure dependent.

biophysics↗

Meta analysis and experimental re-evaluation of the Boyle van 't Hoff relation with osmoregulation modelled by linear elastic principles and ion osmolyte leakage

In this study we challenge the paradigm of using the Boyle van t Hoff (BvH) relation to relate cell size as a linear function of inverse extracellular osmotic pressure for short time periods (~5 to 30 mins). We present alternative models that account for mechanical resistance (turgor model) and ion-osmolyte leakage (leak model), which is not accounted for by the BvH relation. To test the BvH relation and the alternative models, we conducted a meta-analysis of published BvH datasets, as well as new experiments using a HepG2 cell line. Our meta-analysis showed that the BvH relation may be assumed of the hypertonic region but cannot be assumed a priori over the hyper- and hypotonic region. Both alternative models perform better than the BvH relation but are nearly indistinguishable when plotted. The return to isotonic conditions plot indicated neither alternative model accurate predicts return volumes for HepG2 cells. However, a combined turgor-leak model accurately predicts both the BvH plot and the return to isotonic conditions plot. Moreover, this turgor-leak model provides a facile method to estimate the membrane-cortex Youngs modulus and the cell membrane permeability to intracellular ions/osmolytes during periods of osmotic challenge, and predicts a novel passive method of volume regulation without the need for ion pumps.

biophysics↗