Materials Discovery With Machine Learning and Knowledge Discovery

Machine learning and other artificial intelligence methods are gaining increasing prominence in chemistry and materials sciences, especially for materials design and discovery, and in data analysis of results generated by sensors and biosensors. In this paper, we present a perspective on this current use of machine learning, and discuss the prospects of the future impact of extending the use of machine learning to encompass knowledge discovery as an essential step towards a new paradigm of machine-generated knowledge. The reasons why results so far have been limited are given with a discussion of the limitations of machine learning in tasks requiring interpretation. Also discussed is the need to adapt the training of students and scientists in chemistry and materials sciences, to better explore the potential of artificial intelligence capabilities.

On the Challenges for the Diagnosis of SARS-CoV-2 Based on a Review of Current Methodologies

Diagnosis of COVID-19 has been challenging owing to the need for mass testing and for combining distinct types of detection to cover the different stages of the infection. In this review, we have surveyed the most used methodologies for diagnosis of COVID-19, which can be basically categorized into genetic-material detection and immunoassays. Detection of genetic material with real-time polymerase chain reaction (RT-PCR) and similar techniques has been achieved with high accuracy, but these methods are expensive and require time-consuming protocols which are not widely available, especially in less developed countries. Immunoassays for detecting a few antibodies, on the other hand, have been used for rapid, less expensive tests, but their accuracy in diagnosing infected individuals has been limited. We have therefore discussed the strengths and limitations of all of these methodologies, particularly in light of the required combination of tests owing to the long incubation periods. We identified the bottlenecks that prevented mass testing in many countries, and proposed strategies for further action, which are mostly associated with materials science and chemistry. Of special relevance are the methodologies which can be integrated into point-of-care (POC) devices and the use of artificial intelligence that do not require products from a well-developed biotech industry.

O. Microfluidic Point-of-Care Devices: New Trends and Future Prospects for eHealth Diagnostics

Point-of-care (PoC) diagnostics is promising for early detection of a number of diseases, including cancer, diabetes, and cardiovascular diseases, in addition to serving for monitoring health conditions. To be efficient and cost-effective, portable PoC devices are made with microfluidic technologies, with which laboratory analysis can be made with small-volume samples. Recent years have witnessed considerable progress in this area with "epidermal electronics", including miniaturized wearable diagnosis devices. These wearable devices allow for continuous real-time transmission of biological data to the Internet for further processing and transformation into clinical knowledge. Other approaches include bluetooth and WiFi technology for data transmission from portable (non-wearable) diagnosis devices to cellphones or computers, and then to the Internet for communication with centralized healthcare structures. There are, however, considerable challenges to be faced before PoC devices become routine in the clinical practice. For instance, the implementation of this technology requires integration of detection components with other fluid regulatory elements at the microscale, where fluid-flow properties become increasingly controlled by viscous forces rather than inertial forces. Another challenge is to develop new materials for environmentally friendly, cheap, and portable microfluidic devices. In this review paper, we first revisit the progress made in the last few years and discuss trends and strategies for the fabrication of microfluidic devices. Then, we discuss the challenges in lab-on-a-chip biosensing devices, including colorimetric sensors coupled to smartphones, plasmonic sensors, and electronic tongues. The latter ones use statistical and big data analysis for proper classification. The increasing use of big data and artificial intelligence methods is then commented upon in the context of wearable and handled biosensing platforms for the Internet of things and futuristic healthcare systems.

A Future with Ubiquitous Sensing and Intelligent Systems

In this paper, we discuss the relevance of sensing and biosensing for the ongoing revolution in science and technology as a product of the merging of machine learning and Big Data into affordable technologies and accessible everyday products. Possible scenarios for the next decades are described with examples of intelligent systems for various areas, most of which will rely on ubiquitous sensing. The technological and societal challenges for developing the full potential of such intelligent systems are also addressed.

Carbon Nanotube Matrix for Highly Sensitive Biosensors To Detect Pancreatic Cancer Biomarker CA19-9

Biosensors fabricated with nanomaterials promise faster, cheaper, and more efficient alternatives to traditional, often bulky devices for early cancer diagnosis. In this study, we fabricated a thin film sensing unit on interdigitated gold electrodes combining polyethyleneimine and carbon nanotubes in a layer by layer fashion, onto which antibodies anti-CA19-9 were adsorbed with a supporting layer of N-hydroxysuccinimide and 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide solution. By use of impedance spectroscopy, the pancreatic cancer biomarker CA19-9 was detected in a buffer with limit of detection of 0.35 U/mL. This high sensitivity allowed for distinction between samples of blood serum from patients with distinct probabilities to develop pancreatic cancer. The selectivity of the biosensor was confirmed in subsidiary experiments with HT-29 and SW-620 cell lines and possible interferents, e.g., p53 protein, ascorbic acid, and glucose, where significant changes in capacitance could only be measured with HT-29 that contained the CA19-9 biomarker. Chemisorption of CA19-9 molecules onto the layer of anti-CA19-9 antibodies was the mechanism responsible for sensing while electrostatic interactions drove the adsorption of carbon nanotubes, according to polarization-modulated infrared reflection absorption spectroscopy (PM-IRRAS). The adsorption behavior was successfully described by the Langmuir-Freundlich isotherm.

Information Visualization and Feature Selection Methods Applied to Detect Gliadin in Gluten-Containing Foodstuff with a Microfluidic Electronic Tongue

The fast growth of celiac disease diagnosis has sparked the production of gluten-free food and the search for reliable methods to detect gluten in foodstuff. In this paper, we report on a microfluidic electronic tongue (e-tongue) capable of detecting trace amounts of gliadin, a protein of gluten, down to 0.005 mg kg^-1 in ethanol solutions, and distinguishing between gluten-free and gluten-containing foodstuff. In some cases, it is even possible to determine whether gluten-free foodstuff has been contaminated with gliadin. That was made possible with an e-tongue comprising four sensing units, three of which made of layer-by-layer (LbL) films of semiconducting polymers deposited onto gold interdigitated electrodes placed inside microchannels. Impedance spectroscopy was employed as the principle of detection, and the electrical capacitance data collected with the e-tongue were treated with information visualization techniques with feature selection for optimizing performance. The sensing units are disposable to avoid cross-contamination as gliadin adsorbs irreversibly onto the LbL films according to polarization-modulated infrared reflection absorption spectroscopy (PM-IRRAS) analysis. Small amounts of material are required to produce the nanostructured films, however, and the e-tongue methodology is promising for low-cost, reliable detection of gliadin and other gluten constituents in foodstuff.

On the convergence of nanotechnology and Big Data analysis for computer-aided diagnosis

An overview is provided of the challenges involved in building computer-aided diagnosis systems capable of precise medical diagnostics based on integration and interpretation of data from different sources and formats. The availability of massive amounts of data and computational methods associated with the Big Data paradigm has brought hope that such systems may soon be available in routine clinical practices, which is not the case today. We focus on visual and machine learning analysis of medical data acquired with varied nanotech-based techniques and on methods for Big Data infrastructure. Because diagnosis is essentially a classification task, we address the machine learning techniques with supervised and unsupervised classification, making a critical assessment of the progress already made in the medical field and the prospects for the near future. We also advocate that successful computer-aided diagnosis requires a merge of methods and concepts from nanotechnology and Big Data analysis.

Adsorption according to the Langmuir–Freundlich model is the detection mechanism of the antigen p53 for early diagnosis of cancer

The Langmuir–Freundlich model is used to explain the adsorption of the p53 biomarker onto an immunosensor for early detection of cancer.

Controlled Film Architectures to Detect a Biomarker for Pancreatic Cancer Using Impedance Spectroscopy

The need for analytical devices for detecting cancer at early stages has motivated research into nanomaterials where synergy is sought to achieve high sensitivity and selectivity in low-cost biosensors. In this study, we developed a film architecture combining self-assembled monolayer (SAM) and layer-by-layer (LbL) films of polysaccharide chitosan and the protein concanavalin A, on which a layer of anti-CA19-9 antibody was adsorbed. Using impedance spectroscopy with this biosensor, we were capable of detecting low concentrations of the antigen CA19-9, an important biomarker for pancreatic cancer. The limit of detection of 0.69U/mL reached is sufficient for detecting pancreatic cancer at very early stages. The selectivity of the biosensor was inferred from a series of control experiments with samples of cell lines that were tested positive (HT29) and negative (SW620) for the biomarker CA19-9, in addition to the lack of changes in the capacitance value for other analytes and antigen that are not related to this type of cancer. The high sensitivity and selectivity are ascribed to the very specific antigen-antibody interaction, which was confirmed with PM-IRRAS and atomic force microscopy. Also significant is that used information visualization methods to show that different cell lines and commercial samples containing distinct concentrations of CA19-9 and other analytes can be easily distinguished from each other. These computational methods are generic and may be used in optimization procedures to tailor biosensors for specific purposes, as we demonstrated here by comparing the performance of two film architectures in which the concentration of chitosan was varied.