Search for: All records

Raman spectroscopy is a common identification and analysis technique used in research and manufacturing industries. This study investigates the use of Raman spectroscopy and deep learning techniques for identifying various nanofabrication chemicals. Four solvents and SU-8 developer were identified inside common chemical storage and distribution containers. The containers attenuated the spectra and contributed varying amounts of background fluorescence, making manual identification difficult. Two varieties of SU-8 photoresist were differentiated inside amber glass jars, and cured samples of three ratios of polydimethylsiloxane (PDMS) were differentiated using Raman microscopy. The neural network accurately identified the nanofabrication chemicals 100% of the time, without additional preprocessing. This investigation demonstrates the use of Raman spectroscopy and neural networks for the identification of nanofabrication chemicals and makes recommendations for use in other challenging identification applications. 

Bacteria identification can be a time-consuming process. Machine learning algorithms that use deep convolutional neural networks (CNNs) provide a promising alternative. Here, we present a deep learning based approach paired with Raman spectroscopy to rapidly and accurately detect the identity of a bacteria class. We propose a simple 4-layer CNN architecture and use a 30-class bacteria isolate dataset for training and testing. We achieve an identification accuracy of around 86% with identification speeds close to real-time. This optical/biological detection method is promising for applications in the detection of microbes in liquid biopsies and concentrated environmental liquid samples, where fast and accurate detection is crucial. This study uses a recently published dataset of Raman spectra from bacteria samples and an improved CNN model built with TensorFlow. Results show improved identification accuracy and reduced network complexity. 

The use of conventional in vitro and preclinical animal models often fail to properly recapitulate the complex nature of human diseases and hamper the success of translational therapies in humans [1-3] Consequently, research has moved towards organ-on-chip technology to better mimic human tissue interfaces and organ functionality. Herein, we describe a novel approach for the fabrication of a biocompatible membrane made of porous silicon (PSi) for use in organ-on-chip technology that provides key advantages when modeling complex tissue interfaces seen in vivo. By combining well-established methods in the semiconductor industry with organ-on-chip technology, we have developed a novel way of producing thin (25 μm) freestanding PSi biocompatible membranes with both nano (~15.5 nm diameter pores) and macroporous (~0.5 μm diameter pores) structures. To validate the proposed novel membrane, we chose to recapitulate the dynamic environment of the alveolar blood gas exchange interface in alveolar co-culture. Viability assays and immunofluorescence imaging indicate that human pulmonary cells remain viable on the PSi membrane during long-term culture (14 days). Interestingly, it was observed that macrophages can significantly remodel and degrade the PSi membrane substrate in culture. This degradation will allow for more intimate physiological cellular contact between cells, mimicking a true blood-gas exchange interface as observed in vivo. Broadly, we believe that this novel PSi membrane may be used in more complex organ-on-chip and lab-on-chip model systems to accurately recapitulate human anatomy and physiology to provide further insight into human disease pathology and pre-clinical response to therapeutics. 

This is a proof-of-concept study for the development of a field-deployable and low-cost PCR thermocycler (FLC-PCR) to perform Polymerase Chain Reaction (PCR) for the rapid detection of environmental E. coli. Four efficient (77.1 W) peltier modules are used as the central temperature control unit. One 250 W silicone heating pad is used for the heating lid. The PID (proportional-integral-derivative) control algorithm for the thermocycles is implemented by a low-cost 8-bit, 16 MHz microcontroller (ATMEGA328P-PU). ybbW and uidA genes from specific E. coli colonies were used as amplicons for the PCR reactions that were carried out by a commercial PCR machine (Bio-Rad) and our FLC-PCR thermocycler. The heating and cooling speeds averaged 1.11 ± 0.33°C/s which is on a par with the commercial bench-top PCR thermocycler and the efficiency of the heating lid outperformed the Bio-Rad PCR thermocycler. The overall cost of the system is lower than $200 which is more than ten times lower than commercially available units. The heating block can be customized to accommodate different PCR tubes and even microfluidic chambers. An 8000 W portable power generator will be used as the power supply for field studies.