DSpace@RPI (Rensselaer Polytechnic Institute)
Not a member yet
6809 research outputs found
Sort by
Bacterial cellulose : atomized nutrient delivery and gelatin for morphological control
No full textDecember 2021School of ScienceBacterial cellulose (BC) is a naturally derived three-dimensional mesoporous matrix, having high crystallinity, high mechanical strength, high water holding capacity, low density, low electrical conductivity, and is biocompatible. These traits are highly desirable for industrial applications in textiles, air-water filtration, acoustic diaphragms, external wound healing bandages, and tissue scaffolds. The potential benefits of BC use in a wide variety of industrial applications are challenged by long growth times, low cellulose productivity, random fiber distribution, and small average pore size. One way to combat these challenges is to use intermittent top-feeding as a strategy to increase the amount of available nutrients to the bacteria along with assisting in delaying the drop in pH of the growth media from cellulose production, thereby increasing overall BC production. A second way to combat these challenges is the use of non-nutritional additives. These force the bacteria to grow cellulose around areas of high stiffness or template like structures within the media, forming a more organized and larger pore sized matrix. Here, we set out to develop an alternative intermittent top-feeding strategy and study the use of non-nutritional additives on pore formation. First, we developed a method of an intermittent top-feeding strategy utilizing an atomization nozzle apparatus on the growth of BC pellicles. In this method, BC pellicles were fed at 12-hour intervals over a 90-hour cultivations. Here, we found that, by implementing this low-impact top-feeding method, there was a 50% increase in the overall cellulose production by the bacteria compared to normally grown static cultures where all nutrients were provided in the cultivation medium. Pellicles formed by top feeding of atomized nutrient droplets have higher specific modulus and specific tensile strengths then those grown statically. In addition, atomized pellicles had a 250% increase in the water holding capacity along with a higher average porosity. Next, we set out to understand how non-nutritional, naturally derived additives affect the static growth and average pore size development of bacterial cellulose. The additive chosen was gelatin in concentrations of 0.1 wt%, 2.5 wt%, 5 wt%, and 7.5 wt%; as it is a biologically produced material that forms a three-dimensional semi-ordered gel with good thermal stability. We found that, with the addition of gelatin, there is an almost four times increase in cellulose productivity and an almost ten times increase in the average pore size found within the BC matrix. Additionally, there are improved specific modulus and specific tensile strengths seen the gelatin based samples along with a maximum increase of 300% in water holding capacity. In response to these findings, a method was devised to combine the effects of gelatin with the intermittent top-feeding atomization strategy. This was investigated with interval feeding times of 6 hours, 12 hours, and 24 hours over a 7-day period. Additionally, the same total gelatin concentrations used in the static work was implemented in the gelatin media intermittent top-feeding atomization strategy allowing direct comparison of these cultivation strategies. This work revealed that a unique layering morphology was developed within the BC matrix, in which two distinct structural bands were created. The primary structure band resembled that seen in non-gelatin atomized cultures, with larger than natural BC pore sizes and low interstitial fiber counts. The secondary structure band more closely resembled the morphology of the static gelatin pellicles, with maximum average pore sizes of over 20 µm and increasingly thick distinct layer thickness. Overall, the result of this work examined the ability of BC to be fine-tuned for individual needs in industrial applications including the average pore size, productivity, water holding capacity, and mechanical properties. Additionally, our findings informs our understanding of interactions between intermittent top-feeding strategies and non-nutritional viscous additives.Ph
Protein-coated particle capture within a membrane during filtration : simulations
No full textDecember 2020School of EngineeringImproving selectivity between different particles during membrane microfiltration has been a seminal goal for the past 100 years since Sartorius commercialized the first synthetic microporous membranes in Germany. This work is a continuation of the recent approach by Sorci et al. of linking membrane microstructure with filtration performance (Sorci et al., 2019). While they used 2D computational fluid and particle drag mechanics with particle and membrane force measurements described by classical DLVO theory in aqueous solutions, we (i) extend their simulations to 3D using a fluid mechanics simulation program for 3D particle fluid dynamics called MFiX (Multiphase Flow with Interphase eXchanges), and (ii) adapt the extended-DLVO (xDLVO) theory to model the behavior of a plethora of intermolecular force-distance curves between membranes (modified poly(ether sulfone), mPES) and polystyrene particles coated with covalently attached streptavidin at different pH values and salt concentrations. These intermolecular force-distance measurements (including short-term attraction) were obtained using atomic force microscopy (AFM) in force mode by Dr. Mirco Sorci in Dr. Georges Belfort’s research group. The simulation results of particle capture efficiency qualitatively correlated with the magnitude of the short-term attractive forces with a model membrane internal structure comprising an array of spheres. Thus, stronger attractive forces lead to higher particle capture. We also demonstrate, for the first time, that the xDLVO theory describes these jump-in attractive forces, given the solution conditions and net charge on the particle.M
Comprehensive deep learning pipeline for whale shark recognition
No full textMay 2022School of ScienceThe whale shark is the largest fish species in existence today. The main threat to the whale shark population is poaching. Despite conservation efforts, whale shark hunting persists in tropical countries due to population increase and, as a result, growing demand for food. The long maturation period and slow rate of reproduction add to the whale shark population's vulnerability. Whale sharks are listed as endangered species by the International Union for Conservation of Nature, which estimates a 50% decline in the whale shark population over the last 75 years.
Whale sharks migrate over great distances in search of plankton. To date, little is known about whale sharks' life cycle, characteristics of their behavior, and reproduction. Recognition of whale sharks is a starting point for studying the migrations of these animals. In this work, we present an approach for whale shark recognition through a region of interest detection, spot segmentation, and deep metric learning. Whale sharks are speckled with dazzling white spots and lines. Such natural markings are distinctive which makes it possible to achieve good recognition results with modern deep learning techniques.
In this work, we employ a multi-stage approach to tackle the problem of whale shark recognition. Firstly, we prepare a novel whale shark detection dataset and train the YOLOv5s model to detect areas from the pectoral fin to the dorsal fin. This area contains a large amount of whale shark biometric information such as uniquely patterned white spots. Secondly, we train a U-net model with the SEResNet34 backbone to segment these spots on whale sharks' bodies. Thirdly, we train an InceptionResNet embedding model which makes use of spots location as well as originally detected whale shark image to produce high-quality embedding. Finally, we introduce an embedding-based recognition algorithm and validate its performance. For the experiment without new individuals in the test set, our algorithm scores 93% top-1 recognition accuracy, while for the experiment with new individuals in the test set, it scores 83%.M
Cell free production of isobutanol
No full textAugust 2022School of EngineeringWith a need for greener fuels, research into production of biofuels is essential. Isobutanol out preforms ethanol in key metrics such as engine compatibility, energy density, and gasoline blending. Current biofuel strategies of fermentation are constrained by the inherent toxicity of alcohol on microbial cells. While work has been performed on engineering these strains for higher tolerance, cell-free production with enzymes offers a novel approach to bypass the toxicity limitations altogether. These enzymes can also be immobilized to retain enzyme activity and facilitate separations. Based on previous work in the Belfort laboratory, the ketoisovaleric acid pathway was chosen for production of the biofuel, isobutanol. High preforming and stable enzymes were selected from the literature, cloned, expressed, and purified and tested for activity, kinetics, and stability. They were utilized in a novel in vivo to in vitro system, resulting isobutanol titer of 1.78 g/L and yield of 93%. An epoxy immobilized reaction scheme resulted in a titer of 2 g/L and 43% yield. The pathway enzymes were then fused to dockerins, which bound to a cohesin scaffold on cellulose. The reaction utilizing this immobilization scheme resulted in a titer of 5.92 g/L and 78.4% yield. Further work can be done to optimize this reaction, as well as to expand the pathway or scaffold, and incorporate separation of the isobutanol for eventual scaleup.Ph
Chemobiocatalytic Synthesis of a Low Molecular Weight Heparin
No full textACS Chemical Biology,17, 637-646Note : if this item contains full text it may be a preprint, author manuscript, or a Gold OA copy that permits redistribution with a license such as CC BY. The final version is available through the publisher’s platform.Heparin products are widely used clinical anticoagulants essential in the practice of modern medicine. Low-molecular-weight heparins (LMWHs) are currently prepared by the controlled chemical or enzymatic depolymerization of unfractionated heparins (UFHs) that are extracted from animal tissues. In many clinical applications, LMWHs have displaced UFHs and currently comprise over 60% of the heparin market. In the past, our laboratory has made extensive efforts to prepare bioengineered UFHs relying on a chemoenzymatic process to address concerns about animal-sourced UFHs. The current study describes the use of a novel chemoenzymatic process to prepare a chemobiosynthetic LMWH from a low-molecular-weight heparosan. The resulting chemobiocatalytic LMWH matches most of the United States pharmacopeial specifications for enoxaparin, a LMWH prepared through the base-catalyzed depolymerization of animal-derived UFH.National Science Foundationhttps://login.libproxy.rpi.edu/login?url=https://doi.org/10.1021/acschembio.1c0092
Mean-field approaches for network inference and learning
No full textMay 2022School of ScienceMany natural phenomenon such as human activities, wild species evolution and epidemic spreading, man-made complex systems like the World Wide Web, the power grids and the transportation networks
can all be modeled as a networked system
and characterized as a network,
in which the interaction between nodes are governed by a highly coupling and nonlinear dynamics.
The network representation facilitates the understandings of the inner structure and interactions,
hence advances to exercise control over the systems for desired behaviors.
However, the incomplete systems are the norm in practice.
Some components of the systems may be absent due to data inaccessibility,
or corrupted by stochastic noise.
To reconstruct the entire system from incomplete information,
it often involves several related and challenging sub-tasks,
such as recovering the full network (e.g., link prediction),
identifying the explicit dynamics, predicting the steady-states
(e.g., infection rates in epidemic systems, biomass in biological systems),
estimating the topological statistics (e.g., individual degrees, average degree or network motif)
of the complete network, or revealing some crucial behaviors (e.g., traffic breakdown, abrupt blackout)
exhibited during evolution. This dissertation develops a collection of mean-field based approaches, and concentrates on addressing some of these related sub-tasks.
With incomplete information,
we approach reliable inferences of
(i) the equilibrium states and (ii) the individual nodal degrees for general networks,
(iii) the complete structure of the nuclear reaction network and
(iv) the edge dynamics during training for artificial neural networks.
This research studies various complex networks and nonlinear governing dynamics
based on the fundamental mean-field theory.
From an initial state, the system can iteratively evolve into a steady-state.
The steady-state may describe the abundances of different species in a biological system,
or infection rates of different communities in a disease spreading system
when the entire system arrives into an equilibrium point.
The predictability of the steady-state is highly demanding in determining possible human intervention,
especially when the state is undesired or even disastrous to the system.
The research maps the dynamics of the unseen part of the network to a single node,
and it allows us to recover accurate estimates of steady-state on as few as five observed vertices
in domains ranging from ecology to social networks to gene regulation. Another critical part in understanding the dynamical behavior of a complex system is to obtain the full characteristics (e.g., the network size, the degree distribution, the average degree)
of the network from observed data.
Prior studies usually refer to the structure-based estimation,
little effort attempts to estimate the specific degree of each vertex from a sampled induced graph,
which prevents us from measuring the lethality of nodes in protein networks and influencers in social networks.
The current approaches dramatically fail for a tiny sampled induced graph and require a specific sampling method and a large sample size. These approaches neglect information of the vertex state, representing the dynamical behavior of the networked system, such as the biomass of species or expression of a gene, which is useful for degree estimation.
This research fills this gap by integrating the mean-field theory with combinatorial optimization,
and infers individual vertex degrees with both information of the sampled topology and nodes' states.
Experimental results on a variety of real systems demonstrate
that the framework can produce reliable degree estimates and dramatically improve existing link prediction methods by replacing the sampled degrees with the proposed estimates of the degrees. The third task is primarily on the nuclear reaction network. It assembles from a set of existing nuclear reactions a reaction network, whose
degree distribution is found to be bimodal.
That significantly deviates from the common power-law distribution of scale-free networks and
Poisson distribution of random networks.
The research develops a parametric spatial degree model to capture the bimodality,
and proposes a network growth mechanism with three rules to model the structural evolution.
Under the framework, the full reaction network can be reconstructed
by filling the missing links with possible new reactions not yet discovered. Finally, this research brings out a comprehensive analysis of artificial neural networks.The ultimate goal is an efficient model ranking,
identifying a robust neural network model from a set of candidates.
It is a fundamental yet challenging task in deep learning.
Current practice often requires expensive computational costs in training for performance prediction.
This research builds a linear graph representation of a neural network,
then reformulates the stochastic gradient descend based training algorithm to an edge dynamics,
modeling the interactions between synaptic connections.
The analysis is built on the fact that back-propagation during neural network training
is equivalent to the dynamical evolution over the synaptic connections.
Therefore, a converged neural network is associated
with a steady-state of a networked system composed of those edges.
Furthermore, a neural capacitance metric is derived from the edge dynamics as a predictive measure,
universally capturing the performance of the neural network
by observing only a short segment of the early learning curve.
Extensive experiments on a set of popular pre-trained ImageNet models and five benchmark datasets
show that the proposed approach is effective and outperforms the state-of-the-art with as few as five observation points on the learning curves.Ph
One-Pot Enzymatic Synthesis of Heparin from N-Sulfoheparosan
No full textMethods in Molecular Biology, 2303, 3-11Note : if this item contains full text it may be a preprint, author manuscript, or a Gold OA copy that permits redistribution with a license such as CC BY. The final version is available through the publisher’s platform.Heparin, a glycosaminoglycan-based anticoagulant drug, is prepared as an extract of animal tissues. Heparosan, an Escherichia coli (E. coli) K5 capsular polysaccharide with the structure →4)-β-D-glucuronic acid (1 → 4)-β-D-N-acetylglucosamine (1→, corresponds to the precursor backbone in the Golgi-based biosynthesis of heparin. Anticoagulant heparin is prepared in a one-pot synthesis using a chemically prepared derivative of heparosan called N-sulfoheparosan (NSH), recombinant Golgi enzymes expressed in E. coli, and the 3-phosphoadenosine-5-phosphosulfate (PAPS) cofactor.National Institutes of Healthhttps://login.libproxy.rpi.edu/login?url=https://doi.org/10.1007/978-1-0716-1398-6_
Exploring cavity effects on protein dynamic disorder with pressure perturbation
No full textAugust 2022School of ScienceGiven the central role of conformational dynamics in protein function, it is essential to characterize the timescales and structures associated with these transitions. High-pressure perturbation favors transitions to excited states because they typically occupy a smaller molar volume, thus high pressure facilitates the characterization of conformational dynamics. In this dissertation, we describe the use of a combination of NMR chemical exchange spectroscopy, small-angle X-ray scattering, and high hydrostatic pressure to better investigate conformational exchange during protein folding process. Repeat proteins, with their straightforward architecture, provide good models for probing the sequence dependence of protein conformational dynamics. We choose the leucine rich repeat (LRR) domain of the tumor suppressor pp32 as a model. Pp32 is composed of five LRRs with a capping motif on each of its termini. We show here that the introduction of a cavity in the N-terminal capping motif of pp32 leads to pressure-dependent conformational exchange detected on the 500 µs - 2 ms timescale by 15N CPMG relaxation dispersion analysis. Exchange amplitude and minimum chemical shifts decrease from the N- to the C-terminus, revealing a gradient of structural disruption across the protein. In contrast, introduction of a cavity in the central core of pp32 leads to pressure-induced exchange on a slower (> 2 ms) timescale detected by 15N-CEST analysis. Excited state 15N chemical shifts indicate that in the major excited state, the N-terminal region is mostly unfolded, while the core retains native-like structure. These high-pressure chemical state exchange measurements reveal that cavity position dictates distinct structural dynamics, highlighting the subtle, yet central role of sequence in determining protein conformational dynamics.Ph
Drug discovery for Alzheimer's disease targeting the transmembrane domain of amyloid precursor protein (AAPTM).
No full textAugust 2022School of ScienceAlzheimer’s Disease (AD) is a neurological disease currently affecting close to 6 million Americans (3). A major neuropathological hallmark of AD is the presence of senile plaques in the cerebral cortex and hippocampus (3). Senile plaques (amyloid plaques) are mainly composed of extracellular aggregates of amyloid β-peptides (Aβs); it has been hypothesized that Aβ deposition initiates a pathological cascade resulting in cognitive decline characteristic of AD (9-11). γ-secretase (GS) cleaves amyloid precursor protein (APP) in its transmembrane domain generating Aβ40 and Aβ42 that aggregate into the insoluble aggregates of senile plaques (5,11). GS has been a target of anti-AD drug discovery projects with limited success (23-25). Two broad spectrum GS inhibitors (GSIs), avagacestat and semagacestat, failed due to worsening cognition in patients in addition to other serious adverse effects largely attributable to the 90 endogenous substrates of GS (23-25). Despite GSIs’ failure, there is compelling evidence that Aβ is a causative agent in AD, including but not limited to: human genetics of familial AD (FAD) (12-14) and Down’s syndrome (30-31). To circumvent issues from GS inhibition we aimed to target the GS substrate in the amyloidogenic pathway, the transmembrane domain of APP (APPTM). Multiple screening methods resulted in two different binders of APPTM that inhibited APPTM cleavage by GS or presenilin homolog (PSH), 6H8 and N1. 6H8, a covalent modifier of APPTM found in a fragment library, modifies C-terminal lysines in APPTM via a Michael addition mechanism. While N1, a non-covalent modifier found by DNA encoded library screen, binds to APPTM though both hydrophobic and electrostatic interactions. 6H8 and N1 inhibit cleavage of APPTM with IC50 values in the low micromolar and tens of nanomolar ranges, respectively. 6H8 may be engineered into a targeted covalent inhibitor while N1, with nanomolar efficacy, is a promising lead compound for lowering amyloid load for managing AD.Ph