OpenFold: retraining AlphaFold2 yields new insights into its learning mechanisms and capacity for generalization.

AlphaFold2 revolutionized structural biology with the ability to predict protein structures with exceptionally high accuracy. Its implementation, however, lacks the code and data required to train new models. These are necessary to (1) tackle new tasks, like protein-ligand complex structure prediction, (2) investigate the process by which the model learns and (3) assess the model's capacity to generalize to unseen regions of fold space. Here we report OpenFold, a fast, memory efficient and trainable implementation of AlphaFold2. We train OpenFold from scratch, matching the accuracy of AlphaFold2. Having established parity, we find that OpenFold is remarkably robust at generalizing even when the size and diversity of its training set is deliberately limited, including near-complete elisions of classes of secondary structure elements. By analyzing intermediate structures produced during training, we also gain insights into the hierarchical manner in which OpenFold learns to fold. In sum, our studies demonstrate the power and utility of OpenFold, which we believe will prove to be a crucial resource for the protein modeling community.

Mapping variant effects on anti-tumor hallmarks of primary human T cells with base-editing screens.

Single-nucleotide variants (SNVs) in key T cell genes can drive clinical pathologies and could be repurposed to improve cellular cancer immunotherapies. Here, we perform massively parallel base-editing screens to generate thousands of variants at gene loci annotated with known or potential clinical relevance. We discover a broad landscape of putative gain-of-function (GOF) and loss-of-function (LOF) mutations, including in PIK3CD and the gene encoding its regulatory subunit, PIK3R1, LCK, SOS1, AKT1 and RHOA. Base editing of PIK3CD and PIK3R1 variants in T cells with an engineered T cell receptor specific to a melanoma epitope or in different generations of CD19 chimeric antigen receptor (CAR) T cells demonstrates that discovered GOF variants, but not LOF or silent mutation controls, enhanced signaling, cytokine production and lysis of cognate melanoma and leukemia cell models, respectively. Additionally, we show that generations of CD19 CAR T cells engineered with PIK3CD GOF mutations demonstrate enhanced antigen-specific signaling, cytokine production and leukemia cell killing, including when benchmarked against other recent strategies.

Growing ecosystem of deep learning methods for modeling protein-protein interactions.

Numerous cellular functions rely on protein-protein interactions. Efforts to comprehensively characterize them remain challenged however by the diversity of molecular recognition mechanisms employed within the proteome. Deep learning has emerged as a promising approach for tackling this problem by exploiting both experimental data and basic biophysical knowledge about protein interactions. Here, we review the growing ecosystem of deep learning methods for modeling protein interactions, highlighting the diversity of these biophysically informed models and their respective trade-offs. We discuss recent successes in using representation learning to capture complex features pertinent to predicting protein interactions and interaction sites, geometric deep learning to reason over protein structures and predict complex structures, and generative modeling to design de novo protein assemblies. We also outline some of the outstanding challenges and promising new directions. Opportunities abound to discover novel interactions, elucidate their physical mechanisms, and engineer binders to modulate their functions using deep learning and, ultimately, unravel how protein interactions orchestrate complex cellular behaviors.

Nephroprotective Efficacy of Echinops spinosus against a Glycerol-Induced Acute Kidney Injury Model.

Nephroprotection or renal rescue is to revive and restore kidney function after damage, with no need for further dialysis. During acute kidney injury (AKI), sudden and recent reductions in kidney functions occur. Causes are multiple, and prompt intervention can be critical to diminish or prevent morbidity. Echinops spinosus (ES) is a curative plant with proven pharmacological and biological effects including anti-inflammatory, antioxidant, and antibacterial competencies. The principal goal of this research is to scrutinize the nephroprotective features of E. spinosa extract (ESE) against glycerol-induced AKI. Male Wistar albino rats were equally divided into five separated groups: negative control rats (vehicle-injected), ESE control rats (ESE-treated rats), positive control rats, glycerol-induced AKI-model rats (single IM injection of 50% glycerol), and 2 groups of diseased rats but pretreated with different concentrations of ESE for 7 days (ESE150 + AKI rats and ESE250 + AKI rats). Kidney tissues were collected and used for histopathology analysis. The relative kidney weight percentage was assessed. ESE effects were investigated via scanning several biomarkers, such as serum urea and creatinine, as kidney function biomarkers. Lactate dehydrogenase (LDH) and creatine kinase (CK) activities were examined as rhabdomyolysis (RM) indicators. Kidney injury molecule-1 (Kim-1) and neutrophil gelatinase-associated lipocalin (NGAL) were also examined to investigate kidney injury. Enzymatic and nonenzymatic oxidative stress markers were analyzed, namely, superoxide dismutase (SOD), catalase (CAT), glutathione reductase (GR), glutathione peroxidase (GPx), malondialdehyde (MDA), nitric oxide (NO), and reduced glutathione GSH. Proinflammatory cytokine [tumor necrosis factor-α (TNF-α) and interleukin-1 ß (IL-1ß)] and the renal proapoptotic protein (Bax) and antiapoptotic protein (Bcl-2) levels were evaluated. Statistical analysis for the resulting data revealed that ESE pretreatment turned AKI-induced biological antioxidant levels to an extent comparable to normal results. Furthermore, ESE decreased kidney function markers and RM-related biomarkers (LDH, CK, Kim-1, and NGAL) compared to those in untreated AKI-model rats. ESE treatment dropped the apoptotic renal Bax levels, enhanced antiapoptotic Bcl-2 manufacture, and disallowed the release of IL-1ß and TNF-α. This study revealed the protective effect of ESE as therapeutic medicine against AKI-encouraged oxidative stress, inflammation, and apoptosis. It can be effectively used as adjuvant therapy, helping in renal rescue, and for kidney healing in cases with risk factors of AKI.

Association of Oral and General Health with Nutritional Status of Older Adults Attending Two Medical Centers in Riyadh, Saudi Arabia: A Cross-Sectional Study.

Malnutrition could profoundly affect older adults' oral health and quality of life, whereas oral health might, in turn, impact dietary intake and nutritional status. The present study aimed to investigate the association between general and oral health and nutritional status among older adults attending nutrition clinics at two main medical centers in Riyadh, Saudi Arabia. A cross-section study was carried out among adult patients (≥60 years) who attended a geriatric clinic or nutrition clinic at King Khalid University Hospital or King Abdulaziz Medical City, Riyadh. A validated clinician's Mini Nutritional Assessment Short-Form (MNA-SF), Oral Health Impact Profile-5 (OHIP-5), and 36-Item Short Form Survey (SF-36) were collected from each participant. A total of 261 participants with a mean age of 72.14 (±8.97) years were recruited. Diabetes (71%) and hypertension (80%) were present in the majority of patients. The overall MNA-SF score was (10 ± 3). Based on the categorization of the MNA-SF score, 65.9% were classified as malnourished or at risk of malnutrition. Participants with OHIP-5 scores higher than the median (>5) were more likely to be malnourished than those with scores at or lower than 5 (p < 0). The adjusted odd ratio for the MNA-SF score categories indicated that for a one-unit increase in the total SF-36 score, the odds of the malnourished category are 0.94 times less than the risk of malnutrition and normal nutritional status, with OR 0.97 (95% CI 0.94-0.95). Malnutrition or being at risk of malnutrition is likely associated with poor general and oral health. Healthcare providers need to incorporate dietitians into care plans to promote the nutritional health of older adults.

OpenProteinSet: Training data for structural biology at scale.

Multiple sequence alignments (MSAs) of proteins encode rich biological information and have been workhorses in bioinformatic methods for tasks like protein design and protein structure prediction for decades. Recent breakthroughs like AlphaFold2 that use transformers to attend directly over large quantities of raw MSAs have reaffirmed their importance. Generation of MSAs is highly computationally intensive, however, and no datasets comparable to those used to train AlphaFold2 have been made available to the research community, hindering progress in machine learning for proteins. To remedy this problem, we introduce OpenProteinSet, an open-source corpus of more than 16 million MSAs, associated structural homologs from the Protein Data Bank, and AlphaFold2 protein structure predictions. We have previously demonstrated the utility of OpenProteinSet by successfully retraining AlphaFold2 on it. We expect OpenProteinSet to be broadly useful as training and validation data for 1) diverse tasks focused on protein structure, function, and design and 2) large-scale multimodal machine learning research.

Effect of Vitamin D3 on Depressive Behaviors of Rats Exposed to Chronic Unpredictable Mild Stress.

Depression is a psychiatric disorder that negatively affects how a person feels, thinks, and acts. Several studies have reported a positive association between vitamin D (VD) deficiency and depression. Therefore, we aimed to examine the effects of intraperitoneal injection of VD3, fluoxetine (antidepressant), and a combination of VD3 + fluoxetine on a rat model of chronic unpredictable mild stress (CUMS). A total of 40 male Wistar rats (224-296 g) were divided into five groups (n = 8 each) as follows: (1) the control group, (2) the CUMS group, (3) the CUMS group that received vitamin D (10 µg/kg), (4) the CUMS group that received fluoxetine (5 mg/kg), and (5) the CUMS group that received both vitamin D (10 µg/kg) and fluoxetine (5 mg/kg). The CUMS model was produced by exposing rats to frequent social and physical stressors for 21 days. In addition, blood samples were collected to determine corticosterone and serum VD levels. Also, behavioral tests were conducted, including the sucrose preference test (SPT), the forced swimming test (FST), the tail suspension test (TST), the open field test (OFT), and the elevated plus maze test (EPM). Our results show that VD3 had effects similar to fluoxetine on the depressive behavior of the rats when measured by three behavioral tests, namely SPT, FST, and OFT (p < 0.001). Additionally, VD3 had a protective effect against depression similar to that of fluoxetine. Corticosterone levels were lower in the CUMS group that received vitamin D and the CUMS group that received both vitamin D and fluoxetine than in the CUMS group (p < 0.000). In conclusion, VD3 has a protective effect against anxiety and depressive behaviors produced by CUMS in rats.

A Hybrid Structure-Based Machine Learning Approach for Predicting Kinase Inhibition by Small Molecules.

Kinases have been the focus of drug discovery programs for three decades leading to over 70 therapeutic kinase inhibitors and biophysical affinity measurements for over 130,000 kinase-compound pairs. Nonetheless, the precise target spectrum for many kinases remains only partly understood. In this study, we describe a computational approach to unlocking qualitative and quantitative kinome-wide binding measurements for structure-based machine learning. Our study has three components: (i) a Kinase Inhibitor Complex (KinCo) data set comprising in silico predicted kinase structures paired with experimental binding constants, (ii) a machine learning loss function that integrates qualitative and quantitative data for model training, and (iii) a structure-based machine learning model trained on KinCo. We show that our approach outperforms methods trained on crystal structures alone in predicting binary and quantitative kinase-compound interaction affinities; relative to structure-free methods, our approach also captures known kinase biochemistry and more successfully generalizes to distant kinase sequences and compound scaffolds.

High-throughput deep learning variant effect prediction with Sequence UNET.

Understanding coding mutations is important for many applications in biology and medicine but the vast mutation space makes comprehensive experimental characterisation impossible. Current predictors are often computationally intensive and difficult to scale, including recent deep learning models. We introduce Sequence UNET, a highly scalable deep learning architecture that classifies and predicts variant frequency from sequence alone using multi-scale representations from a fully convolutional compression/expansion architecture. It achieves comparable pathogenicity prediction to recent methods. We demonstrate scalability by analysing 8.3B variants in 904,134 proteins detected through large-scale proteomics. Sequence UNET runs on modest hardware with a simple Python package.

Massively parallel base editing screens to map variant effects on anti-tumor hallmarks of primary human T cells.

Base editing enables generation of single nucleotide variants, but large-scale screening in primary human T cells is limited due to low editing efficiency, among other challenges 1 . Here, we developed a high-throughput approach for high-efficiency and massively parallel adenine and cytosine base-editor screening in primary human T cells. We performed multiple large-scale screens editing 102 genes with central functions in T cells and full-length tiling mutagenesis of selected genes, and read out variant effects on hallmarks of T cell anti-tumor immunity, including activation, proliferation, and cytokine production. We discovered a broad landscape of gain- and loss-of-function mutations, including in PIK3CD and its regulatory subunit encoded by PIK3R1, LCK , AKT1, CTLA-4 and JAK1 . We identified variants that affected several (e.g., PIK3CD C416R) or only selected (e.g. LCK Y505C) hallmarks of T cell activity, and functionally validated several hits by probing downstream signaling nodes and testing their impact on T cell polyfunctionality and proliferation. Using primary human T cells in which we engineered a T cell receptor (TCR) specific to a commonly presented tumor testis antigen as a model for cellular immunotherapy, we demonstrate that base edits identified in our screens can tune specific or broad T cell functions and ultimately improve tumor elimination while exerting minimal off-target activity. In summary, we present the first large-scale base editing screen in primary human T cells and provide a framework for scalable and targeted base editing at high efficiency. Coupled with multi-modal phenotypic mapping, we accurately nominate variants that produce a desirable T cell state and leverage these synthetic proteins to improve models of cellular cancer immunotherapies.

Single-sequence protein structure prediction using a language model and deep learning.

AlphaFold2 and related computational systems predict protein structure using deep learning and co-evolutionary relationships encoded in multiple sequence alignments (MSAs). Despite high prediction accuracy achieved by these systems, challenges remain in (1) prediction of orphan and rapidly evolving proteins for which an MSA cannot be generated; (2) rapid exploration of designed structures; and (3) understanding the rules governing spontaneous polypeptide folding in solution. Here we report development of an end-to-end differentiable recurrent geometric network (RGN) that uses a protein language model (AminoBERT) to learn latent structural information from unaligned proteins. A linked geometric module compactly represents Cα backbone geometry in a translationally and rotationally invariant way. On average, RGN2 outperforms AlphaFold2 and RoseTTAFold on orphan proteins and classes of designed proteins while achieving up to a 106-fold reduction in compute time. These findings demonstrate the practical and theoretical strengths of protein language models relative to MSAs in structure prediction.

Diagnostic Value of Inflammatory Markers in Elderly Arab Women with Sarcopenia.

Several studies have found a correlation between inflammatory markers and sarcopenia; however, limited research has been conducted on the Arabic population. Therefore, this study aimed to investigate the value of inflammatory parameters in Saudi elderly women with sarcopenia. In this cross-sectional study, 76 elderly Saudi women (>65 years) were stratified according to the presence (n = 26) or absence (n = 50) of sarcopenia, using the operational definition of the Asian Working Group for Sarcopenia (AWGS). Demographics and clinical data were collected. Muscle mass, muscle strength, and physical performance were assessed using bioelectrical impedance, hand grip and timed-up-and-go (TUG) tests, respectively. Inflammatory markers such as interleukin-6 (IL-6), tumor necrosis factor alpha (TNF-α) and C-reactive protein (CRP) were assessed using commercially available assays. Muscle mass and strength indicators were lower in the sarcopenia group (p-value < 0.05). Moreover, interleukin 6 (IL-6) was positively correlated with TUG (r = 0.48, p-value < 0.05), while CRP showed an inverse correlation with the right leg muscle (R-Leg-M) and a positive correlation with triceps skinfold (TSF) (r = −0.41, r = 0.42, respectively, p-values < 0.05). Additionally, TSF and R-Leg-M were independent predictors of CRP variation (R2 = 0.35; p < 0.01). Lastly, participants with a TNF-α > 71.2 were five times more likely to have sarcopenia [(OR = 5.85), 95% CI: 1.07−32.08; p = 0.04]. In conclusion, elevated levels of TNF-α are significantly associated with the risk of sarcopenia, while variations perceived in circulating CRP can be explained by changes in the muscle mass indices only among individuals with sarcopenia. The present findings, while promising, need further investigations on a larger scale to determine whether inflammatory markers hold any diagnostic value in assessing sarcopenia among elderly Arab women.

The next decade of protein structure.

With recent dramatic advances in various techniques used for protein structure research, we asked researchers to comment on the next exciting questions for the field and about how these techniques will advance our knowledge not only about proteins but also about human health and diseases.

Podocyte specific deletion of PKM2 ameliorates LPS-induced podocyte injury through beta-catenin.

BACKGROUND: Acute kidney injury (AKI) is associated with a severe decline in kidney function caused by abnormalities within the podocytes' glomerular matrix. Recently, AKI has been linked to alterations in glycolysis and the activity of glycolytic enzymes, including pyruvate kinase M2 (PKM2). However, the contribution of this enzyme to AKI remains largely unexplored. METHODS: Cre-loxP technology was used to examine the effects of PKM2 specific deletion in podocytes on the activation status of key signaling pathways involved in the pathophysiology of AKI by lipopolysaccharides (LPS). In addition, we used lentiviral shRNA to generate murine podocytes deficient in PKM2 and investigated the molecular mechanisms mediating PKM2 actions in vitro. RESULTS: Specific PKM2 deletion in podocytes ameliorated LPS-induced protein excretion and alleviated LPS-induced alterations in blood urea nitrogen and serum albumin levels. In addition, PKM2 deletion in podocytes alleviated LPS-induced structural and morphological alterations to the tubules and to the brush borders. At the molecular level, PKM2 deficiency in podocytes suppressed LPS-induced inflammation and apoptosis. In vitro, PKM2 knockdown in murine podocytes diminished LPS-induced apoptosis. These effects were concomitant with a reduction in LPS-induced activation of ß-catenin and the loss of Wilms' Tumor 1 (WT1) and nephrin. Notably, the overexpression of a constitutively active mutant of ß-catenin abolished the protective effect of PKM2 knockdown. Conversely, PKM2 knockdown cells reconstituted with the phosphotyrosine binding-deficient PKM2 mutant (K433E) recapitulated the effect of PKM2 depletion on LPS-induced apoptosis, ß-catenin activation, and reduction in WT1 expression. CONCLUSIONS: Taken together, our data demonstrates that PKM2 plays a key role in podocyte injury and suggests that targetting PKM2 in podocytes could serve as a promising therapeutic strategy for AKI. TRIAL REGISTRATION: Not applicable. Video abstract.

Value of Adding Bioelectrical Impedance Analysis to Anthropometric Indices in the Diagnosis of Metabolic Syndrome in 10-16 Years Old Schoolgirls.

The use of bioelectrical impedance analysis (BIA) in clinical settings is common. However, the value of BIA-based parameters in diagnosing metabolic syndrome (MetS) in children is under-investigated. Herein, we aimed to study the usefulness of BIA-indices in the diagnoses of MetS in 6-10-year-old girls. Therefore, a diagnostic accuracy case-control study was conducted, which included 75 girls aged 10-16 years, divided into three age-matched groups (normal, None-MetS, and MetS). Anthropometric indices, BIA parameters (including fat-free mass (FFM), body fat percent (BFP), and total body water (TBW)), blood pressure (BP), and blood samples were collected. Our main findings show that for girls in None-MetS and MetS groups, the waist circumference (WC) correlated positively with waist-hip ratio and mid-arm circumference (r = 0.58, 0.47, respectively), but not with BFP based on skinfold thickness (SFT), or mid-arm muscle area. WC was positively correlated with FFM and TBW, while high-density lipoprotein was inversely correlated with FFM. However, fasting blood glucose, triglycerides and BP showed no association with anthropometric measurements and BIA components. WC was the best indicator of MetS (AUC = 0.88, cut-off = 81.5 cm), followed by BMI (AUC = 0.84, cut-off = 26.9 kg/m2), while BFP based on SFT was the least sensitive (62.5%). In conclusion, apart from the FM index, anthropometric parameters such as WC are more valuable in diagnosing MetS in young adolescent girls.

Differentiable biology: using deep learning for biophysics-based and data-driven modeling of molecular mechanisms.

Deep learning using neural networks relies on a class of machine-learnable models constructed using 'differentiable programs'. These programs can combine mathematical equations specific to a particular domain of natural science with general-purpose, machine-learnable components trained on experimental data. Such programs are having a growing impact on molecular and cellular biology. In this Perspective, we describe an emerging 'differentiable biology' in which phenomena ranging from the small and specific (for example, one experimental assay) to the broad and complex (for example, protein folding) can be modeled effectively and efficiently, often by exploiting knowledge about basic natural phenomena to overcome the limitations of sparse, incomplete and noisy data. By distilling differentiable biology into a small set of conceptual primitives and illustrative vignettes, we show how it can help to address long-standing challenges in integrating multimodal data from diverse experiments across biological scales. This promises to benefit fields as diverse as biophysics and functional genomics.

Protein structure prediction by AlphaFold2: are attention and symmetries all you need?

The functions of most proteins result from their 3D structures, but determining their structures experimentally remains a challenge, despite steady advances in crystallography, NMR and single-particle cryoEM. Computationally predicting the structure of a protein from its primary sequence has long been a grand challenge in bioinformatics, intimately connected with understanding protein chemistry and dynamics. Recent advances in deep learning, combined with the availability of genomic data for inferring co-evolutionary patterns, provide a new approach to protein structure prediction that is complementary to longstanding physics-based approaches. The outstanding performance of AlphaFold2 in the recent Critical Assessment of protein Structure Prediction (CASP14) experiment demonstrates the remarkable power of deep learning in structure prediction. In this perspective, we focus on the key features of AlphaFold2, including its use of (i) attention mechanisms and Transformers to capture long-range dependencies, (ii) symmetry principles to facilitate reasoning over protein structures in three dimensions and (iii) end-to-end differentiability as a unifying framework for learning from protein data. The rules of protein folding are ultimately encoded in the physical principles that underpin it; to conclude, the implications of having a powerful computational model for structure prediction that does not explicitly rely on those principles are discussed.