Motivation

Problems with current methods of detection

Fluorescence spectroscopy, used in detection methods (with aptamers and antibodies) have a 5% error margin. Processing the fluorescence data further adds to this error. Problems with microscopy techniques involve loss of information which we call “cell-to-pixel” loss. This is because the number of possible pixels is limited in a screen area and they do not fully capture biological matter or events. Thus conversion of biological signals/entities into digital signals leads to increase in noise and information loss.

Another most common method of bacterial/fungal infection is using cultures. Most disease-causing bacteria require one to two days of culture, but some types of bacteria take five days or longer to grow enough cells. Fungal cultures, aka fungal smears, are more difficult, and it may take weeks to identify the cause of fungal infection.

A faster and more accurate method of detection is using PCR techniques (such as colony PCR). However, this requires a highly trained professional, which is seldom available in remote areas of third-world countries. The COVID-19 pandemic is an example of how testing numbers were low in developing nations like India due to the limited availability of equipment and personnel for PCR.

Antibody-based test kits such as ELISA and the Rapid Antigen Test developed for COVID-19 suffer from one fatal flaw-high false positives or false negatives. Antibodies depend on the immunogenicity of the pathogen (the ability of the pathogen to induce an immune response), which varies from microbe to microbe. Antibodies, being proteins, are very sensitive to pH, temperature, and contaminants. Furthermore, the production of antibodies is an expensive and time-consuming process.

Increasing bacterial and fungal infections

Bacterial infections have been plaguing the world forever. Septicemia (sepsis) is a fatal organ dysfunction in response to primarily bacterial (can also be fungal or viral infection) in the blood that currently has a 17% mortality rate, which jumps to 26% in severe cases in our country. In addition, due to antimicrobial resistance, there is an ever-increasing trend of patients getting infected by multi-drug-resistant bacteria.

Escherichia coli, one of the ESKAPE, has multiple strains pathogenic to humans and animals. E. coli is the cause of urinary tract infections, especially prevalent in rural areas with poor sanitation. Diarrheagenic E. coli is responsible for gastrointestinal infections, which are fatal in infants. Various strains of E. coli are also becoming antibiotic-resistant, which is a cause for concern.

Salmonella typhimurium and closely related species like Salmonella typhi are the leading cause of gastroenteric infections accompanied by diarrhea. Due to rampant use of antibiotics in poultry farms, fisheries, cattle industry, there is an ever-rising population of multi-drug resistant Salmonella causing fatal infections in humans as well as livestock.

The tropical climate, malnutrition, and unhygienic living situations also cause an alarming number of lethal fungal infections. The detection of fungi is difficult, and research concerning new antifungals is slow, resulting in 32.55 out of 1000 patients being sent to the Intensive Care Unit with fungal infections in the national capital.

Food and Water contamination

In our region and the country as a whole, the majority of the population, being low-income, is exposed to less-than-ideal sanitary conditions, contaminated food, and poor-quality potable water. According to WHO, 110 billion USD is lost yearly in productivity and medical expenses resulting from unsafe food in low- and middle-income countries like India. Our location being Punjab, an agricultural, dairy, and food industry hub, gives us first-hand exposure to these problems. Thus, we have chosen the bacteria E. coli and S. typhi and the fungus Penicillium all of which commonly contaminate food and water.

Sustainable computing

Current electronics and computers use integrated circuit chips that have silicon as an important component. Silicon/silica mining is a highly polluting process, leading to deforestation, soil erosion, and health hazards like silicosis for the miners.

Graphical Processing Units (GPUs) and High-Performance Clusters (HPCs) are used for implementing machine learning algorithms, AI, bitcoin mining, data analysis, etc., resulting in a huge carbon footprint and heat generation. Even the personal computers we use are responsible for carbon and heat emissions. This indirectly contributes to global warming and climate change.

The disposal of electronic waste (e-waste) is a huge problem. Improper disposal of electronics leads to soil contamination and water contamination due to the leaching of toxic metals like gold (Au), silver (Ag), mercury(Hg), cadmium(Cd), etc.

In developing countries such as India, there is an exponential increase in the use of computers and electronic devices, causing an increase in the aforementioned problems- pollution due to silicon mining, an increase in carbon footprint, and soil/water pollution due to improper e-waste disposal.

Our Solution

We have developed a platform called NeuraSyn incorporating a neural chip and aptamers and demonstrating the system by detection of microbes - bacteria Escherichia coli and Salmonella typhimurium and fungal product Penicillium, produced by Penicllium chrysogenum. We have created a neural chip - a biological computer made of trained neurons that can detect electrical signals and take decisions. We use aptamers which bind to the target, causing an increase in impedance which the neural chip senses and decides which microbes are present.

NeuraSyn uses a neural chip that interacts directly with the aptamers leading to a minimization of information loss. It also amplifies the input signals leading to increased accuracy in decision making. This would make our system less prone to false positives (like fluorescent-based detection systems).

Our platform aims to be a zero-time detection kit which is reusable and eco-friendly (the detection entities are biodegradable, non-toxic entities).

The Neural Chip

Neurons are the computational cells in an organism. They are highly sensitive to changes in the electrical activity of the external environment and therefore are perfect for making devices that require low-power detection capabilities. We hence decided to use these properties of neurons to detect the changes in the electrical impedance of Aptamers. This would make our system less prone to false positives (like fluorescent-based detection systems), make simultaneous multi-detection possible and bypass the errors from spectrophotometer detection and image processing cell to pixel losses.

The neural chip works by attaching differentiated SH-SY 5Y cells and N2A cells to make two prototypes of chips. Our system contained a chip, with input and output electrodes sputtered using a layer of chromium and topped with Gold. This made the system conductive, durable and biosafe. The center of the chip contains a well, made from a falcon tube. This gives us space for growing the differentiated cells directly on the chip, reducing transferring steps. We attached a filter on top of the falcon to allow for gas exchange in the well but prevent contamination.

The working of the chip was similar to that of an Ensemble of Neurons connected to an echo state network (ESN), of sparse connection. The ESN is a random connection of excitable nodes. The action potential fired by these nodes is computed by the trained network. The training of the network is dependent on the hebbian plasticity rules.

Where w_ij is the connection strength between the ith dendrite and the jth axon terminal, and x refers to the product of the currents.

The short term memory capacity of the neural chip can be mathematically expressed as

Where C is the memory capacity and r²(u(n-i),y_i(n)), is the squared correlation coefficient between the input signal.

The biological connectivity of the cells could be seen by microscopy.

The cells differentiated on the palate showed neural morphology by day 8 and could be seen connected to the electrodes. The cells are the main detection and processing system in our project. The input electrode brings in an electrical signal from the aptamers. This is then sensed by the neurons. Depending on the input current into the system different numbers of neurons are activated. The number of neurons that therefore fire an action potential is proportional to the amount of input current. Hence a decrease in the current, indicative of an increase in the aptamer bound to the target, is a signal of the pathogen being present. In the detection system, we are only concerned with the electrical nature and the trainability of the neuron-like cells. This helps us evolve the circuit for the detection and processing of the signal within the system just by changing the input parameters.

Aptamers

Aptamers are oligonucleotides or short strands of DNA or RNA. Aptamers have the property of binding to specific small molecules (drugs, vitamins, metal ions, etc.), macromolecules (proteins, glycoproteins, lipoproteins, nucleic acids, etc.), viruses, and even whole cells like bacteria, via hydrogen bonding, electrostatic interactions, hydrophobic interactions (Van der Waals interactions), π-stacking interactions, etc.

Aptamers have tremendous potential in biosensing (detection of substances) and therapeutics due to the following properties:

1. Highly sensitive - detects very few target molecules in the sample.
2. Highly specific - aptamers show very little cross-reactivity, they interact only with the target molecules/cells.
3. Independent of immunogenicity - aptamers, unlike antibody-based detection, aptamers do not depend on the body’s immune response to a pathogen, as they bind directly to the pathogen
4. Low cost, increased stability, reduced batch invariably, and high tunability (can be modified easily)

Aptamers against target molecules are produced by the Systemic Evolution of Ligands by EXponential enrichment (SELEX). For the detection of substances, electrochemical methods, fluorescence, and optical methods can be used. We use the principle of increase in impedance when aptamers bind to the target (electrochemical method) for our platform NeuraSyn.

We have identified three aptamers against bacteria Escherichia coli ATCC 25922 and Salmonella typhimurium, and fungal product Penicillin G (Benzyl penicillin) from Penicillium chrysogenum through an extensive literature search. Experiments performed with the aptamers may be summarized as follows:

1. We placed the aptamer on a GO/PANI (Graphene Oxide and Polyaniline sheet) on a glassy carbon electrode.
2. We then introduced the bacteria/target molecules
3. We measured the Electrochemical Impedance Spectroscopy (EIS) and Cyclic Voltammetry to identify the impedance increase when the aptamers bind to the target molecules.

Electrochemical impedance spectroscopy:

Electrochemical impedance spectroscopy (EIS) is a technique used for the analysis of interfacial properties related to bio-recognition events occurring at the electrode surface. It measures impedance in a circuit is in ohms (resistance unit). Electrochemical processes associated with the electrolyte/interface and redox reactions are simulated/computed as an electric circuit (equivalent circuit) involving electrical components (resistors, capacitors, inductors). This equivalent circuit is designed and implemented to understand and evaluate the individual components of the EIS system.

Resistance of solution (Rs), double layer capacitance at the surface of the electrode (CdI), charge transfer resistance (Rct), and Warburg resistance (Zw) are simplified in the Randles equivalent circuits. EIS generates a Nyquist plot showing the impedance changes at different frequencies.

Cyclic Voltammetry:

Cyclic voltammetry is an electrochemical technique for measuring the current response of a redox active solution to a linearly cycled potential sweep between two or more set values. It is a useful method for quickly determining information about the thermodynamics of redox processes, the energy levels of the analyte and the kinetics of electronic-transfer reactions.

It uses a three electrode system, like EIS, consisting of a working electrode, reference electrode, and counter or auxiliary electrode.

To perform cyclic voltammetry, the electrolyte solution is first added to an electrochemical cell along with a reference solution and the three electrodes. A potentiostat is then used to linearly sweep the potential between the working and reference electrodes until it reaches a pre-set limit, at which point it is swept back in the opposite direction.

This process is repeated multiple times during a scan and the changing current between the working and counter probes is measured by the device in real time. The result is a characteristic duck-shaped plot known as a cyclic voltammogram.