An incremental learning approach to prediction models of SEIRD variables in the context of the COVID-19 pandemic.

Several works have proposed predictive models of the SEIRD (Susceptible, Exposed, Infected, Recovered, and Dead) variables to characterize the pandemic of COVID-19. One of the challenges of these models is to be able to follow the dynamics of the disease to make more precise predictions. In this paper, we propose an approach based on incremental learning to build predictive models of the SEIRD variables for the COVID-19 pandemic. Our incremental learning approach is a dynamic ensemble method based on a bagging scheme that allows the addition of new models or the updating of incremental models. The article proposes an incremental learning architecture composed of two components. The first component carries out an analysis of the interdependencies of the SEIRD variables and the second component is an incremental learning model that builds/updates the predictive models. The paper analyses the quality of the predictive models of our incremental learning approach using data of the COVID-19 from Colombia, and shows interesting results about the predictions of the SEIRD variables. These results are compared with an incremental learning approach based on random forests.

Modulating smooth muscle cell response by the release of TGFß2 from tubular scaffolds for vascular tissue engineering.

Tissue engineering has gained considerable attention in the development of small diameter tissue engineered vascular grafts (TEVGs) for treating coronary heart disease. A properly designed acellular and biodegradable TEVG must encourage the infiltration and growth of vascular smooth muscle cells (SMCs). Our group has previously shown that increasing levels of TGFß2 can differentially modulate SMC migration and proliferation. In this study, tubular electrospun scaffolds loaded with TGFß2 were fabricated using various ratios of gelatin/polycaprolactone (PCL), resulting in scaffolds with porous nano-woven architecture suitable for tissue ingrowth. Scaffold morphology, degradation rate, TGß2 release kinetics, and bioactivity were assessed. TGFß2 was successfully integrated into the electrospun biomaterial that resulted in a differential release profile depending on the gelatin/PCL ratio over the course of 42 days. Higher TGFß2 elution was obtained in scaffolds with higher gelatin content, which may be related to the biodegradation of gelatin in culture media. The biological activity of the released TGFß2 was evaluated by its ability to affect SMC proliferation as a function of its concentration. SMCs seeded on TGFß2-loaded scaffolds also showed higher densities and infiltration after 5 days in culture as compared to scaffolds without TGFß2. Our results demonstrate that the ratio of synthetic and natural polymers in electrospun blends can be used to tune the release of TGFß2. This method can be used to intelligently modulate the SMC response in gelatin/PCL scaffolds making the TGFß2-loaded conduits attractive for cardiovascular tissue engineering applications.

2-Photon Characterization of Optical Proteolytic Beacons for Imaging Changes in Matrix-Metalloprotease Activity in a Mouse Model of Aneurysm.

Abdominal aortic aneurysm is a multifactorial disease that is a leading cause of death in developed countries. Matrix-metalloproteases (MMPs) are part of the disease process, however, assessing their role in disease initiation and progression has been difficult and animal models have become essential. Combining Förster resonance energy transfer (FRET) proteolytic beacons activated in the presence of MMPs with 2-photon microscopy allows for a novel method of evaluating MMP activity within the extracellular matrix (ECM). Single and 2-photon spectra for proteolytic beacons were determined in vitro. Ex vivo experiments using the apolipoprotein E knockout angiotensin II-infused mouse model of aneurysm imaged ECM architecture simultaneously with the MMP-activated FRET beacons. 2-photon spectra of the two-color proteolytic beacons showed peaks for the individual fluorophores that enable imaging of MMP activity through proteolytic cleavage. Ex vivo imaging of the beacons within the ECM revealed both microstructure and MMP activity. 2-photon imaging of the beacons in aneurysmal tissue showed an increase in proteolytic cleavage within the ECM (p<0.001), thus indicating an increase in MMP activity. Our data suggest that FRET-based proteolytic beacons show promise in assessing MMP activity within the ECM and will therefore allow future studies to identify the heterogeneous distribution of simultaneous ECM remodeling and protease activity in aneurysmal disease.

Biomechanical Comparison of Glutaraldehyde-Crosslinked Gelatin Fibrinogen Electrospun Scaffolds to Porcine Coronary Arteries.

Cardiovascular disease (CVD) is the leading cause of death for Americans. As coronary artery bypass graft surgery (CABG) remains a mainstay of therapy for CVD and native vein grafts are limited by issues of supply and lifespan, an effective readily available tissue-engineered vascular graft (TEVG) for use in CABG would provide drastic improvements in patient care. Biomechanical mismatch between vascular grafts and native vasculature has been shown to be the major cause of graft failure, and therefore, there is need for compliance-matched biocompatible TEVGs for clinical implantation. The current study investigates the biaxial mechanical characterization of acellular electrospun glutaraldehyde (GLUT) vapor-crosslinked gelatin/fibrinogen cylindrical constructs, using a custom-made microbiaxial optomechanical device (MOD). Constructs crosslinked for 2, 8, and 24 hrs are compared to mechanically characterized porcine left anterior descending coronary (LADC) artery. The mechanical response data were used for constitutive modeling using a modified Fung strain energy equation. The results showed that constructs crosslinked for 2 and 8 hrs exhibited circumferential and axial tangential moduli (ATM) similar to that of the LADC. Furthermore, the 8-hrs experimental group was the only one to compliance-match the LADC, with compliance values of 0.0006±0.00018 mm Hg-1 and 0.00071±0.00027 mm Hg-1, respectively. The results of this study show the feasibility of meeting mechanical specifications expected of native arteries through manipulating GLUT vapor crosslinking time. The comprehensive mechanical characterization of cylindrical biopolymer constructs in this study is an important first step to successfully develop a biopolymer compliance-matched TEVG.

Semiconductor laser with extended cavity and intracavity atomic filter.

We report on peculiar dynamic features of laser oscillation in a cavity with a semiconductor junction as the gain medium and an intracavity atomic absorber. The output face of the semiconductor is antireflection coated, and lasing action is achieved by using a diffraction grating to close the laser cavity. The spectral analysis of the laser emission evidences a stable emission with narrow linewidth when the oscillating frequency is resonant with the atomic absorber. We also observe frequency bistability and instability. The change between these regimes is controlled through the bias current in a very reproducible way.