Paneth cell dysfunction in radiation injury and radio-mitigation by human α-defensin 5.

Introduction: The mechanism underlying radiation-induced gut microbiota dysbiosis is undefined. This study examined the effect of radiation on the intestinal Paneth cell α-defensin expression and its impact on microbiota composition and mucosal tissue injury and evaluated the radio-mitigative effect of human α-defensin 5 (HD5). Methods: Adult mice were subjected to total body irradiation, and Paneth cell α-defensin expression was evaluated by measuring α-defensin mRNA by RT-PCR and α-defensin peptide levels by mass spectrometry. Vascular-to-luminal flux of FITC-inulin was measured to evaluate intestinal mucosal permeability and endotoxemia by measuring plasma lipopolysaccharide. HD5 was administered in a liquid diet 24 hours before or after irradiation. Gut microbiota was analyzed by 16S rRNA sequencing. Intestinal epithelial junctions were analyzed by immunofluorescence confocal microscopy and mucosal inflammatory response by cytokine expression. Systemic inflammation was evaluated by measuring plasma cytokine levels. Results: Ionizing radiation reduced the Paneth cell α-defensin expression and depleted α-defensin peptides in the intestinal lumen. α-Defensin down-regulation was associated with the time-dependent alteration of gut microbiota composition, increased gut permeability, and endotoxemia. Administration of human α-defensin 5 (HD5) in the diet 24 hours before irradiation (prophylactic) significantly blocked radiation-induced gut microbiota dysbiosis, disruption of intestinal epithelial tight junction and adherens junction, mucosal barrier dysfunction, and mucosal inflammatory response. HD5, administered 24 hours after irradiation (treatment), reversed radiation-induced microbiota dysbiosis, tight junction and adherens junction disruption, and barrier dysfunction. Furthermore, HD5 treatment also prevents and reverses radiation-induced endotoxemia and systemic inflammation. Conclusion: These data demonstrate that radiation induces Paneth cell dysfunction in the intestine, and HD5 feeding prevents and mitigates radiation-induced intestinal mucosal injury, endotoxemia, and systemic inflammation.

Evolutionary-scale prediction of atomic-level protein structure with a language model.

Recent advances in machine learning have leveraged evolutionary information in multiple sequence alignments to predict protein structure. We demonstrate direct inference of full atomic-level protein structure from primary sequence using a large language model. As language models of protein sequences are scaled up to 15 billion parameters, an atomic-resolution picture of protein structure emerges in the learned representations. This results in an order-of-magnitude acceleration of high-resolution structure prediction, which enables large-scale structural characterization of metagenomic proteins. We apply this capability to construct the ESM Metagenomic Atlas by predicting structures for >617 million metagenomic protein sequences, including >225 million that are predicted with high confidence, which gives a view into the vast breadth and diversity of natural proteins.

Toward generalizable prediction of antibody thermostability using machine learning on sequence and structure features.

Over the last three decades, the appeal for monoclonal antibodies (mAbs) as therapeutics has been steadily increasing as evident with FDA's recent landmark approval of the 100th mAb. Unlike mAbs that bind to single targets, multispecific biologics (msAbs) have garnered particular interest owing to the advantage of engaging distinct targets. One important modular component of msAbs is the single-chain variable fragment (scFv). Despite the exquisite specificity and affinity of these scFv modules, their relatively poor thermostability often hampers their development as a potential therapeutic drug. In recent years, engineering antibody sequences to enhance their stability by mutations has gained considerable momentum. As experimental methods for antibody engineering are time-intensive, laborious and expensive, computational methods serve as a fast and inexpensive alternative to conventional routes. In this work, we show two machine learning approaches - one with pre-trained language models (PTLM) capturing functional effects of sequence variation, and second, a supervised convolutional neural network (CNN) trained with Rosetta energetic features - to better classify thermostable scFv variants from sequence. Both of these models are trained over temperature-specific data (TS50 measurements) derived from multiple libraries of scFv sequences. On out-of-distribution (refers to the fact that the out-of-distribution sequnes are blind to the algorithm) sequences, we show that a sufficiently simple CNN model performs better than general pre-trained language models trained on diverse protein sequences (average Spearman correlation coefficient, ρ, of 0.4 as opposed to 0.15). On the other hand, an antibody-specific language model performs comparatively better than the CNN model on the same task (ρ= 0.52). Further, we demonstrate that for an independent mAb with available thermal melting temperatures for 20 experimentally characterized thermostable mutations, these models trained on TS50 data could identify 18 residue positions and 5 identical amino-acid mutations showing remarkable generalizability. Our results suggest that such models can be broadly applicable for improving the biological characteristics of antibodies. Further, transferring such models for alternative physicochemical properties of scFvs can have potential applications in optimizing large-scale production and delivery of mAbs or bsAbs.

End-to-end learning of multiple sequence alignments with differentiable Smith-Waterman.

MOTIVATION: Multiple sequence alignments (MSAs) of homologous sequences contain information on structural and functional constraints and their evolutionary histories. Despite their importance for many downstream tasks, such as structure prediction, MSA generation is often treated as a separate pre-processing step, without any guidance from the application it will be used for. RESULTS: Here, we implement a smooth and differentiable version of the Smith-Waterman pairwise alignment algorithm that enables jointly learning an MSA and a downstream machine learning system in an end-to-end fashion. To demonstrate its utility, we introduce SMURF (Smooth Markov Unaligned Random Field), a new method that jointly learns an alignment and the parameters of a Markov Random Field for unsupervised contact prediction. We find that SMURF learns MSAs that mildly improve contact prediction on a diverse set of protein and RNA families. As a proof of concept, we demonstrate that by connecting our differentiable alignment module to AlphaFold2 and maximizing predicted confidence, we can learn MSAs that improve structure predictions over the initial MSAs. Interestingly, the alignments that improve AlphaFold predictions are self-inconsistent and can be viewed as adversarial. This work highlights the potential of differentiable dynamic programming to improve neural network pipelines that rely on an alignment and the potential dangers of optimizing predictions of protein sequences with methods that are not fully understood. AVAILABILITY AND IMPLEMENTATION: Our code and examples are available at: https://github.com/spetti/SMURF. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Interpreting Potts and Transformer Protein Models Through the Lens of Simplified Attention.

The established approach to unsupervised protein contact prediction estimates coevolving positions using undirected graphical models. This approach trains a Potts model on a Multiple Sequence Alignment. Increasingly large Transformers are being pretrained on unlabeled, unaligned protein sequence databases and showing competitive performance on protein contact prediction. We argue that attention is a principled model of protein interactions, grounded in real properties of protein family data. We introduce an energy-based attention layer, factored attention, which, in a certain limit, recovers a Potts model, and use it to contrast Potts and Transformers. We show that the Transformer leverages hierarchical signal in protein family databases not captured by single-layer models. This raises the exciting possibility for the development of powerful structured models of protein family databases.

Evaluating Protein Transfer Learning with TAPE.

Machine learning applied to protein sequences is an increasingly popular area of research. Semi-supervised learning for proteins has emerged as an important paradigm due to the high cost of acquiring supervised protein labels, but the current literature is fragmented when it comes to datasets and standardized evaluation techniques. To facilitate progress in this field, we introduce the Tasks Assessing Protein Embeddings (TAPE), a set of five biologically relevant semi-supervised learning tasks spread across different domains of protein biology. We curate tasks into specific training, validation, and test splits to ensure that each task tests biologically relevant generalization that transfers to real-life scenarios. We benchmark a range of approaches to semi-supervised protein representation learning, which span recent work as well as canonical sequence learning techniques. We find that self-supervised pretraining is helpful for almost all models on all tasks, more than doubling performance in some cases. Despite this increase, in several cases features learned by self-supervised pretraining still lag behind features extracted by state-of-the-art non-neural techniques. This gap in performance suggests a huge opportunity for innovative architecture design and improved modeling paradigms that better capture the signal in biological sequences. TAPE will help the machine learning community focus effort on scientifically relevant problems. Toward this end, all data and code used to run these experiments are available at https://github.com/songlab-cal/tape.

Human Defensin-5 Blocks Ethanol and Colitis-Induced Dysbiosis, Tight Junction Disruption and Inflammation in Mouse Intestine.

Alcohol consumption has been shown to cause dysbiosis, but the mechanism involved in it is unknown. Recurrent colitis is known to induce expression of α-defensins in the colon, but the effect of alcohol consumption on it is not known. We investigated the effect of ethanol on α-defensin expression in the small intestine and colitis-induced expression in colon in mice. Furthermore, we evaluated the effect of human defensin-5 (HD5) on ethanol and colitis-induced gut barrier dysfunction and mucosal damage. Recurrent colitis was induced by feeding dextran sulfate sodium (DSS), 3 cycles of 5-days each with 15 days intervals, followed by 30-days remission. Ethanol was fed during the intervals and recovery in a liquid diet with or without HD5. Expression of α-defensins, tight junction (TJ) integrity and cytokine/chemokine expression were analyzed. Chronic ethanol feeding reduced α-defensin expression in the small intestine and colitis-induced defensin expression in the colon. HD5 attenuated the growth of enterotoxigenic Bacteriodes fragilis and E. coli, but had no effect on non-toxigenic Bacteriodes fragilis or probiotics, the Lactobacilli. Ethanol and colitis elevated Enterobacteriaceae, Firmicutes and Firmicutes to Bacteriodetes ratio in colonic mucosa. HD5 feeding attenuated ethanol and colitis-induced dysbiosis, disruption of intestinal epithelial TJ, mucosal inflammation, expression of pro-inflammatory cytokines and chemokines in the small intestine and colon, and endotoxemia. These results demonstrate that ethanol suppresses intestinal α-defensin expression, leading to dysbiosis, barrier dysfunction, inflammation and endotoxemia. HD5 feeding attenuates intestinal injury caused by ethanol and colitis, indicating that defensin expression is a potential target for treatment of alcoholic tissue injury and colitis.

An updated review on factors and their inter-linked influences on photovoltaic system performance.

Globally installed solar photovoltaics (PV) capacity has crossed three hundred gigawatts and is increasing each year. As the share of solar PV in the energy mix of a country increases, forecasting PV power available will be crucial. To forecast the instantaneous and long-term PV power output, understanding the factors influencing them is necessary. In this view, this work elaborates on the factors that impact the PV system through tabulation and graphical explanation. Further, a discussion of the articles related to the dust-induced change in performance is made. To understand the impact of dust on solar PV systems in depth, advanced instrumentation and methodologies have been used in the past few years. One of the methods is the measurement of spectral transmittance/reflectance/absorptance of the dust layer on the PV panel. This has led to the question whether a thin layer of some specific dust can be beneficial by absorbing infrared (IR) heat and hence allowing the PV cells to operate at a lower temperature. Many controlled experiments in the laboratory have been made using the artificial dust and sun simulators; and such studies aid in the development of numerical models. Research in modeling, mathematical analysis (from first principles) of dust deposition, and calculation of its impact on panels have been given importance in recent years. Outdoor experiments are relatively more common than other modes of research in this field. Studies involving the interaction of deposited dust with spectral radiation, improving the correlation between artificial and natural dust deposition, the interplay between dust and atmospheric parameters are to be encouraged.

Calcium Channels and Oxidative Stress Mediate a Synergistic Disruption of Tight Junctions by Ethanol and Acetaldehyde in Caco-2 Cell Monolayers.

Ethanol is metabolized into acetaldehyde in most tissues. In this study, we investigated the synergistic effect of ethanol and acetaldehyde on the tight junction integrity in Caco-2 cell monolayers. Expression of alcohol dehydrogenase sensitized Caco-2 cells to ethanol-induced tight junction disruption and barrier dysfunction, whereas aldehyde dehydrogenase attenuated acetaldehyde-induced tight junction disruption. Ethanol up to 150 mM did not affect tight junction integrity or barrier function, but it dose-dependently increased acetaldehyde-mediated tight junction disruption and barrier dysfunction. Src kinase and MLCK inhibitors blocked this synergistic effect of ethanol and acetaldehyde on tight junction. Ethanol and acetaldehyde caused a rapid and synergistic elevation of intracellular calcium. Calcium depletion by BAPTA or Ca2+-free medium blocked ethanol and acetaldehyde-induced barrier dysfunction and tight junction disruption. Diltiazem and selective knockdown of TRPV6 or CaV1.3 channels, by shRNA blocked ethanol and acetaldehyde-induced tight junction disruption and barrier dysfunction. Ethanol and acetaldehyde induced a rapid and synergistic increase in reactive oxygen species by a calcium-dependent mechanism. N-acetyl-L-cysteine and cyclosporine A, blocked ethanol and acetaldehyde-induced barrier dysfunction and tight junction disruption. These results demonstrate that ethanol and acetaldehyde synergistically disrupt tight junctions by a mechanism involving calcium, oxidative stress, Src kinase and MLCK.

Coronary pseudoaneurysm: an uncommon cause of P.U.O. after a common procedure: an interesting case report.

INTRODUCTION: Pyrexia 2 months after a coronary procedure is usually an unrelated event; we came across a rather uncommon cause of delayed appearance of fever related to subacute development of pseudoaneurysm and pericardial effusion secondary to leakage from the pseudoaneurysm after primary angioplasty of LAD, and to the best of our knowledge pseudoaneurysm leak as a cause of prolonged pyrexia is not reported in the literature. BACKGROUND: Two and half months after successful primary angioplasty with stent implantation for an acute myocardial infarction, we were confronted with an unexplained fever of prolonged duration, after extensive work the cause of unexplained pyrexia was found to be related to development of hemorrhagic pericardial effusion secondary to leak from coronary pseudoaneurysm. The pseudoaneurysm was located in ostium of left anterior descending coronary artery. Despite review of the literature and an examination of the procedural data, we were unable to explain the cause of pseudoaneurysm formation. We assume that either the guidewire perforated the arterial wall when it passed the occlusion or a dissection not visualized angiographically, provoked this phenomenon.