Proteins perform many essential functions in biological systems and can be successfully developed as bio-therapeutics. It is invaluable to be able to predict their properties based on a proposed sequence and structure. In this study, we developed a novel generalizable deep learning framework, LM-GVP, composed of a protein Language Model (LM) and Graph Neural Network (GNN) to leverage information from both 1D amino acid sequences and 3D structures of proteins. Our approach outperformed the state-of-the-art protein LMs on a variety of property prediction tasks including fluorescence, protease stability, and protein functions from Gene Ontology (GO). We also illustrated insights into how a GNN prediction head can inform the fine-tuning of protein LMs to better leverage structural information. We envision that our deep learning framework will be generalizable to many protein property prediction problems to greatly accelerate protein engineering and drug development.
Primary dystonia's prolonged muscle contractions and the associated abnormal postures and twisting movements remain incurable. Genetic mutation/deletion of GAG from TorsonA's gene resulting in ΔE303 (which weakens the binding between TorsinA and its activator, such as LULL1) primarily cause this neurodegenerative disorder. We studied TorsinA-LULL1 (or TorsinAΔE303-LULL1) bindings and interactions. For the first time, we show the atomic details of TorsinA-LULL1 dynamic interactions and TorsinAΔE303-LULL1 dynamic interactions and their binding affinities. Our results show extensive effects of ΔE303 on TorsinAΔE303-LULL1 interactions, and suggest that the differences between TorsinA-LULL1 interactions and TorsinAΔE303-LULL1 interactions are non-subtle. ΔE303 significantly weakens TorsinAΔE303-LULL1's binding affinity. We present pieces of evidence proving that the effects of ΔE303 (on the differences between TorsinA-LULL1 interactions and TorsinAΔE303-LULL1 interactions) are more pronounced than previously suggested, and that the nanobody used for achieving the X-ray crystallization in the previous study attenuated the differences between TorsinA-LULL1 and TorsinAΔE303-LULL1 interactions. Our accounts of the dynamic interactions between “TorsinA and LULL1” and between “TorsinAΔE303 and LULL1” and the detailed effects of ΔE303 on TorsinA-/TorsinAΔE303-LULL1 build on previous findings and offer new insights for a better understanding of the molecular basis of Primary Dystonia. Our results have long-term potentials of guiding the development of medications for the disease.
BACKGROUND:Lead, an example of heavy metals, has, for decades, being known for its adverse effects on various body organs and systems such that their functions are compromised.AIM:In the present study, the ability of lead to adversely affect the male reproductive system was investigated and tomato (Lycopersicon esculentum: Source of antioxidants) paste (TP) was administered orally to prevent the adverse effects of Pb.MATERIALS AND METHODS:Fifteen Sprague Dawley rats, randomised into three groups (n = 5), were used for this study. Animals in Group A served as the control and were drinking distilled water. Animals in Groups B and C were drinking 1% Pb (II) acetate (LA). Group C animals were, in addition to drinking LA, treated with 1.5 ml of TP/day. All treatments were for 8 weeks.STATISTICAL ANALYSIS USED:A Mann–Whitney U-test was used to analyse the results obtained.RESULTS:The obtained results showed that Pb caused a significant reduction in the testicular weight, sperm count, life–death ratio, sperm motility, normal sperm morphology, and plasma and tissue superoxide dismutase and catalase activity, but a significant increase in plasma and tissue malondialdehyde concentration. But, Pb did not cause any significant change in the serum testosterone level. TP, however, significantly reduced these adverse effects of Pb.CONCLUSION:These findings lead to the conclusion that TP significantly lowered the adverse effects of Pb exposure on the kidney as well as Pb-induced oxidative stress.
In this (modest) study, we developed artificial neural network (ANN) models for predicting body weight using various independent (input) variables in eight-week old New Zealand white purebred and crossbred rabbits. From the whole data sets of similar age groups, 75 percent were used to train the neural network model and 25 percent were used to test the effectiveness of the model. Five predictor variables were used viz, breed, sex, heart girth, body length and height at wither as input variables and body weight was considered as dependent variable from the model. The ANN used was multilayer feed forward network with back propagation of error for efficient learning. Our ANN models (with R 2 = 0.68 at ten thousand iterations, and R 2 = 0.71 one million iterations) performed better than traditional multivariate linear regression (MLR) models (R 2 = 0.66) indicating that the ANN models were able to more accurately capture how the variations in input variables explained the variations in body weight. It is concluded that ANN models are more powerful than MLR models in predicting animals' body weight. Nonetheless, we recognize that fitting an ANN model requires more computation resources than fitting a tradition MLR model but the benefits of its accuracy outweigh any demerit from the associated computation overhead.
Plasmodium falciparum malaria, which degrades haemoglobin through falcipain-2 (FP2), is a serious disease killing 445 thousand people annually. Since the P. falciparum’s survival in humans depends on its ability to degrade human’s haemoglobin, stoppage or hindrance of FP2 has antimalarial effects. Therefore, we studied the atomic details of how E64 approaches, binds to, and inhibits FP2. We found that E64 (1) gradually approaches FP2 by first interacting with FP2’s D170 and Q171 or N81, N77, and K76; (2) binds FP2 tightly (ΔGbinding = −12.2 ± 1.1 kJ/mol); and (3) persistently blocks access to FP2’s catalytic residues regardless of whether or not E64 has already been able to form a covalent bond with FP2’s C42. Furthermore, the results suggest that S41, D234, D170, N38, N173, and L172 (which are located in or near the FP2’s catalytic site’s binding pocket) contribute the most towards the favourable binding of E64 to FP2. Their in silico mutations adversely affect E64-FP2 binding affinity with D234L/A, N173L/A, W43F/A, D234L/A, H174F/A, and N38L/A having the most significant adverse effects on E64-FP2 binding and interactions. The findings presented in this article, which has antimalarial implications, suggest that hydrogen bonding and electrostatic interactions play important roles in E64-FP2 binding, and that a potential FP2-blocking E64-based/E64-like antimalarial drug should be capable of being both hydrogen-bond donor and acceptor, and/or have the ability to favourably interact with polar amino acids (such as S41, S149, N38, N173, N77, Q171) and with charged amino acids (such as D234, D170, H174) of FP2. The abilities to favourably interact with ASN, ASP, and SER appears to be important characteristics that such potential drug should have.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.