AbDPP: Target-oriented antibody design with pretraining and prior biological structure knowledge.

Antibodies represent a crucial class of complex protein therapeutics and are essential in the treatment of a wide range of human diseases. Traditional antibody discovery methods, such as hybridoma and phage display technologies, suffer from limitations including inefficiency and a restricted exploration of the immense space of potential antibodies. To overcome these limitations, we propose a novel method for generating antibody sequences using deep learning algorithms called AbDPP (target-oriented antibody design with pretraining and prior biological knowledge). AbDPP integrates a pretrained model for antibodies with biological region information, enabling the effective use of vast antibody sequence data and intricate biological system understanding to generate sequences. To target specific antigens, AbDPP incorporates an antibody property evaluation model, which is further optimized based on evaluation results to generate more focused sequences. The efficacy of AbDPP was assessed through multiple experiments, evaluating its ability to generate amino acids, improve neutralization and binding, maintain sequence consistency, and improve sequence diversity. Results demonstrated that AbDPP outperformed other methods in terms of the performance and quality of generated sequences, showcasing its potential to enhance antibody design and screening efficiency. In summary, this study contributes to the field by offering an innovative deep learning-based method for antibody generation, addressing some limitations of traditional approaches, and underscoring the importance of integrating a specific antibody pretrained model and the biological properties of antibodies in generating novel sequences. The code and documentation underlying this article are freely available at https://github.com/zlfyj/AbDPP.

Unbiased curriculum learning enhanced global-local graph neural network for protein thermodynamic stability prediction.

MOTIVATION: Proteins play crucial roles in biological processes, with their functions being closely tied to thermodynamic stability. However, measuring stability changes upon point mutations of amino acid residues using physical methods can be time-consuming. In recent years, several computational methods for protein thermodynamic stability prediction (PTSP) based on deep learning have emerged. Nevertheless, these approaches either overlook the natural topology of protein structures or neglect the inherent noisy samples resulting from theoretical calculation or experimental errors. RESULTS: We propose a novel Global-Local Graph Neural Network powered by Unbiased Curriculum Learning for the PTSP task. Our method first builds a Siamese graph neural network to extract protein features before and after mutation. Since the graph's topological changes stem from local node mutations, we design a local feature transformation module to make the model focus on the mutated site. To address model bias caused by noisy samples, which represent unavoidable errors from physical experiments, we introduce an unbiased curriculum learning method. This approach effectively identifies and re-weights noisy samples during the training process. Extensive experiments demonstrate that our proposed method outperforms advanced protein stability prediction methods, and surpasses state-of-the-art learning methods for regression prediction tasks. AVAILABILITY AND IMPLEMENTATION: All code and data is available at https://github.com/haifangong/UCL-GLGNN.

Theoretical Investigation of the Structural Characteristics in the Active State of Akt1 Kinase.

Akt (known as protein kinase B or PKB) is a serine/threonine kinase that regulates multiple biological processes, including cell growth, survival, and differentiation. Akt plays a critical role in the intracellular signaling network through conformational changes responsive to diverse signal inputs, and dysregulation of Akt activity could give rise to a number of diseases. However, understanding of Akt's dynamic structures and conformational transitions between active and inactive states remains unclear. In this work, classical MD simulations and QM/MM calculations were carried out to unveil the structural characteristics of Akt1, especially in its active state. The doubly protonated H194 was investigated, and both ATP-Akt1 and ADP-Akt1 complexes were constructed to demonstrate the significance of ATP in maintaining the ATP-K179-E198 salt bridge and the corresponding allosteric pathway. Besides, conformational transitions from the inactive state to the active state showed different permeation patterns of water molecules in the ATP pocket. The coordination modes of Mg2+ in the dominant representative conformations ( I and I') are presented. Unlike the water-free conformation I', three water molecules appear around Mg2+ in the water-occupied conformation I, which can finally exert an influence on the catalytic mechanism of Akt1. Furthermore, QM/MM calculations were performed to study the phosphoryl-transfer reaction of Akt1. The transfer of ATP γ-phosphate was achieved through a reversible conformational change from the reactant to a critical prereaction state, with a water molecule moving into the reaction center to coordinate with Mg2+, after which the γ-phosphate group was transferred from ATP to the substrate. Taken together, our results elucidate the structural characteristics of the Akt1 active state and shed new light on the catalytic mechanism of Akt kinases.

Neoantigen load as a prognostic and predictive marker for stage II/III non-small cell lung cancer in Chinese patients.

BACKGROUND: The prognosis for patients with stage II/III non-small cell lung cancer (NSCLC) is unsatisfactory, even after complete tumor resection and adjuvant chemotherapy. Here, we assessed the prognostic and predictive value of immunogenomic signatures for stage II/III NSCLC in Chinese patients. METHODS: A total of 91 paired resected stage II/III NSCLC and normal tissues, including 47 squamous cell lung carcinomas (SCC) and 44 lung adenocarcinomas (ADC), were collected and analyzed using whole exome sequencing (WES) to identify immunogenomic signatures for association with clinicopathological variables and disease-free survival (DFS). RESULTS: Higher neoantigen load (NAL, >2 neoantigens/Mb) exhibited better DFS for SCC patients (p = 0.021) but not ADC patients. A benefit from adjuvant chemotherapy was correlated with lower NAL (≤2 neoantigens/Mb) (p = 0.009). However, tumor mutation burden (TMB), mutations of individual gene, oncogene pathways, and antigen presentation machinery genes, and human leukocyte antigen (HLA)-I number and HLA-I loss of heterozygosity (LOH) had no prognostic or predictive value for DFS of SCC or ADC patients. CONCLUSIONS: NAL is a useful biomarker for lung SCC prognosis and prediction of chemotherapy responses in Chinese patients. The predictive value of NAL for adjuvant immunotherapy should be further explored in patients with resected NSCLC.

Diagnostic performance of convolutional neural network-based Tanner-Whitehouse 3 bone age assessment system.

BACKGROUND: Bone age can reflect the true growth and development status of a child; thus, it plays a critical role in evaluating growth and endocrine disorders. This study established and validated an optimized Tanner-Whitehouse 3 artificial intelligence (TW3-AI) bone age assessment (BAA) system based on a convolutional neural network (CNN). METHODS: A data set of 9,059 clinical radiographs of the left hand was obtained from the picture archives and communication systems (PACS) between January 2012 and December 2016. Among these, 8,005/9,059 (88%) samples were treated as the training set for model implementation, 804/9,059 (9%) samples as the validation set for parameters optimization, and the remaining 250/9,059 (3%) samples were used to verify the accuracy and reliability of the model compared to that of 4 experienced endocrinologists and 2 experienced radiologists. The overall variation of TW3-metacarpophalangeal, radius, ulna and short bones (RUS) and TW3-Carpal bone score, as well as each bone (13 RUS + 7 Carpal) between reviewers and the AI, were compared by Bland-Altman (BA) chart and Kappa test, respectively. Furthermore, the time consumption between the model and reviewers was also compared. RESULTS: The performance of TW3-AI model was highly consistent with the reviewers' overall estimation, and the root mean square (RMS) was 0.50 years. The accuracy of the BAA of the TW3-AI model was better than the estimate of the reviewers. Further analysis revealed that human interpretations of the male capitate, hamate, the first distal and fifth middle phalanx and female capitate, the trapezoid, and the third and fifth middle phalanx, were most inconsistent. The average image processing time was 1.5±0.2 s in the TW3-AI model, which was significantly shorter than manual interpretation. CONCLUSIONS: The diagnostic performance of CNN-based TW3 BAA was accurate and timesaving, and possesses better stability compared to diagnostics made by experienced experts.

Understanding the R882H mutation effects of DNA methyltransferase DNMT3A: a combination of molecular dynamics simulations and QM/MM calculations.

DNA (cytosine-5)-methyltransferase 3A (DNMT3A), a key enzyme for de novo epigenetic methylation in human beings, was reported to undergo an R882H mutation in approximately 25% of M4/M5 subtype acute myeloid leukemia (AML) patients. In this work, a combination of classical molecular dynamics (MD) simulations and QM/MM calculation methods was utilized to reveal the molecular mechanism behind the activity attenuation caused by R882H mutation. We found that R882H mutation induces a "folded" conformation in the methyl donor S-adenosylmethionine (SAM) through different types of hydrogen bond formation at the terminal carbonyl oxygen atom and the hydroxyl O3' atom of the ribose ring on SAM, with Arg891 as a mediator. Energetically, both the pre-reaction state (PRS) and transition state (TS) were stabilized in the R882H mutant. However, the energy barrier of the rate-determining step from the PRS to the TS was calculated to be roughly 1.0 kcal mol-1 larger in the R882H mutant than the WT. Also, a dynamic transformation occurred along the helix where R882H was located, tending to manifest in a quasi-"Newton's cradle" manner from the mutational site to the active site residues of DNMT3A. Our computational results provided molecular insights into the pathogenic R882H mutation and advanced the understanding of its mechanism.

HSP70-Hrd1 axis precludes the oncorepressor potential of N-terminal misfolded Blimp-1s in lymphoma cells.

B lymphocyte-induced maturation protein-1 (Blimp-1) ensures B-cell differentiation into the plasma cell stage, and its instability constitutes a crucial oncogenic element in certain aggressive cases of activated B cell-like diffuse large B-cell lymphoma (ABC-DLBCL). However, the underlying degradation mechanisms and their possible therapeutic relevance remain unexplored. Here, we show that N-terminal misfolding mutations in ABC-DLBCL render Blimp-1 protein susceptible to proteasome-mediated degradation but spare its transcription-regulating activity. Mechanistically, whereas wild-type Blimp-1 metabolism is triggered in the nucleus through PML-mediated sumoylation, the degradation of lymphoma-associated mutants is accelerated by subversion of this pathway to Hrd1-mediated cytoplasmic sequestration and ubiquitination. Screening experiments identifies the heat shock protein 70 (HSP70) that selects Blimp-1 mutants for Hrd1 association, and HSP70 inhibition restores their nuclear accumulation and oncorepressor activities without disrupting normal B-cell maturation. Therefore, HSP70-Hrd1 axis represents a potential therapeutic target for restoring the oncorepressor activity of unstable lymphoma-associated Blimp-1 mutants.The transcriptional repressor Blimp-1 has an important role in B-cell differentiation. Here the authors show that lymphoma-associated Blimp-1 mutants are selectively recognized by HSP70-Hrd1, which leads to their accelerated degradation and propose HSP70 inhibition as a therapeutic approach for certain lymphomas.