Evasive Spike Variants Elucidate the Preservation of T Cell Immune Response to the SARS-CoV-2 Omicron Variant.

The Omicron variants boast the highest infectivity rates among all SARS-CoV-2 variants. Despite their lower disease severity, they can reinfect COVID-19 patients and infect vaccinated individuals as well. The high number of mutations in these variants render them resistant to antibodies that otherwise neutralize the spike protein of the original SARS-CoV-2 spike protein. Recent research has shown that despite its strong immune evasion, Omicron still induces strong T Cell responses similar to the original variant. This work investigates the molecular basis for this observation using the neural network tools NetMHCpan-4.1 and NetMHCiipan-4.0. The antigens presented through the MHC Class I and Class II pathways from all the notable SARS-CoV-2 variants were compared across numerous high frequency HLAs. All variants were observed to have equivalent T cell antigenicity. A novel positive control system was engineered in the form of spike variants that did evade T Cell responses, unlike Omicron. These evasive spike proteins were used to statistically confirm that the Omicron variants did not exhibit lower antigenicity in the MHC pathways. These results suggest that T Cell immunity mounts a strong defense against COVID-19 which is difficult for SARS-CoV-2 to overcome through mere evolution.

Neural network execution using nicked DNA and microfluidics.

DNA has been discussed as a potential medium for data storage. Potentially it could be denser, could consume less energy, and could be more durable than conventional storage media such as hard drives, solid-state storage, and optical media. However, performing computations on the data stored in DNA is a largely unexplored challenge. This paper proposes an integrated circuit (IC) based on microfluidics that can perform complex operations such as artificial neural network (ANN) computation on data stored in DNA. We envision such a system to be suitable for highly dense, throughput-demanding bio-compatible applications such as an intelligent Organ-on-Chip or other biomedical applications that may not be latency-critical. It computes entirely in the molecular domain without converting data to electrical form, making it a form of in-memory computing on DNA. The computation is achieved by topologically modifying DNA strands through the use of enzymes called nickases. A novel scheme is proposed for representing data stochastically through the concentration of the DNA molecules that are nicked at specific sites. The paper provides details of the biochemical design, as well as the design, layout, and operation of the microfluidics device. Benchmarks are reported on the performance of neural network computation.

Computing mathematical functions with chemical reactions via stochastic logic.

This paper presents a novel strategy for computing mathematical functions with molecular reactions, based on theory from the realm of digital design. It demonstrates how to design chemical reaction networks based on truth tables that specify analog functions, computed by stochastic logic. The theory of stochastic logic entails the use of random streams of zeros and ones to represent probabilistic values. A link is made between the representation of random variables with stochastic logic on the one hand, and the representation of variables in molecular systems as the concentration of molecular species, on the other. Research in stochastic logic has demonstrated that many mathematical functions of interest can be computed with simple circuits built with logic gates. This paper presents a general and efficient methodology for translating mathematical functions computed by stochastic logic circuits into chemical reaction networks. Simulations show that the computation performed by the reaction networks is accurate and robust to variations in the reaction rates, within a log-order constraint. Reaction networks are given that compute functions for applications such as image and signal processing, as well as machine learning: arctan, exponential, Bessel, and sinc. An implementation is proposed with a specific experimental chassis: DNA strand displacement with units called DNA "concatemers".

Evaluating NetMHCpan performance on non-European HLA alleles not present in training data.

Bias in neural network model training datasets has been observed to decrease prediction accuracy for groups underrepresented in training data. Thus, investigating the composition of training datasets used in machine learning models with healthcare applications is vital to ensure equity. Two such machine learning models are NetMHCpan-4.1 and NetMHCIIpan-4.0, used to predict antigen binding scores to major histocompatibility complex class I and II molecules, respectively. As antigen presentation is a critical step in mounting the adaptive immune response, previous work has used these or similar predictions models in a broad array of applications, from explaining asymptomatic viral infection to cancer neoantigen prediction. However, these models have also been shown to be biased toward hydrophobic peptides, suggesting the network could also contain other sources of bias. Here, we report the composition of the networks' training datasets are heavily biased toward European Caucasian individuals and against Asian and Pacific Islander individuals. We test the ability of NetMHCpan-4.1 and NetMHCpan-4.0 to distinguish true binders from randomly generated peptides on alleles not included in the training datasets. Unexpectedly, we fail to find evidence that the disparities in training data lead to a meaningful difference in prediction quality for alleles not present in the training data. We attempt to explain this result by mapping the HLA sequence space to determine the sequence diversity of the training dataset. Furthermore, we link the residues which have the greatest impact on NetMHCpan predictions to structural features for three alleles (HLA-A*34:01, HLA-C*04:03, HLA-DRB1*12:02).

Hydrophobicity identifies false positives and false negatives in peptide-MHC binding.

Major Histocompability Complex (MHC) Class I molecules allow cells to present foreign and endogenous peptides to T-Cells so that cells infected by pathogens can be identified and killed. Neural networks tools such as NetMHC-4.0 and NetMHCpan-4.1 are used to predict whether peptides will bind to variants of MHC molecules. These tools are trained on data gathered from binding affinity and eluted ligand experiments. However, these tools do not track hydrophobicity, a significant biochemical factor relevant to peptide binding, in their predictions. A previous study had concluded that the peptides predicted to bind to HLA-A*0201 by NetMHC-4.0 were much more hydrophobic than expected. This paper expands that study by also focusing on HLA-B*2705 and HLA-B*0801, which prefer binding hydrophilic and balanced peptides respectively. The correlation of hydrophobicity of 9-mer peptides with their predicted binding strengths to these various HLAs was investigated. Two studies were performed, one using the data that the two neural networks were trained on, and the other using a sample of the human proteome. NetMHC-4.0 was found to have a statistically significant bias towards predicting highly hydrophobic peptides as strong binders to HLA-A*0201 and HLA-B*2705 in both studies. Machine Learning metrics were used to identify the causes for this bias: hydrophobic false positives and hydrophilic false negatives. These results suggest that the retraining the neural networks with biochemical attributes such as hydrophobicity and better training data could increase the accuracy of their predictions. This would increase their impact in applications such as vaccine design and neoantigen identification.