3D graph contrastive learning for molecular property prediction.

MOTIVATION: Self-supervised learning (SSL) is a method that learns the data representation by utilizing supervision inherent in the data. This learning method is in the spotlight in the drug field, lacking annotated data due to time-consuming and expensive experiments. SSL using enormous unlabeled data has shown excellent performance for molecular property prediction, but a few issues exist. (i) Existing SSL models are large-scale; there is a limitation to implementing SSL where the computing resource is insufficient. (ii) In most cases, they do not utilize 3D structural information for molecular representation learning. The activity of a drug is closely related to the structure of the drug molecule. Nevertheless, most current models do not use 3D information or use it partially. (iii) Previous models that apply contrastive learning to molecules use the augmentation of permuting atoms and bonds. Therefore, molecules having different characteristics can be in the same positive samples. We propose a novel contrastive learning framework, small-scale 3D Graph Contrastive Learning (3DGCL) for molecular property prediction, to solve the above problems. RESULTS: 3DGCL learns the molecular representation by reflecting the molecule's structure through the pretraining process that does not change the semantics of the drug. Using only 1128 samples for pretrain data and 0.5 million model parameters, we achieved state-of-the-art or comparable performance in six benchmark datasets. Extensive experiments demonstrate that 3D structural information based on chemical knowledge is essential to molecular representation learning for property prediction. AVAILABILITY AND IMPLEMENTATION: Data and codes are available in https://github.com/moonkisung/3DGCL.

Clinicopathological Characteristics of PIK3CA Mutation and Amplification in Korean Patients with Breast Cancers.

The frequency of PIK3CA mutation and amplification was various and their clinical significances have not been clarified in Korean patients with invasive breast carcinoma (IBC). The study aimed to investigate the clinical and prognostic significances of PIK3CA mutation and amplification in IBC patients. DNA was isolated from paired normal and tumoral tissues in 128 IBC patients and the mutation and expression of PIK3CA gene were analyzed. PIK3CA mutation and expression was detected in 14.3% and 21.9% of IBC patients, respectively. And the level of PIK3CA expression was not different according to the presence of PIK3CA mutation (p = 0.775). PIK3CA mutation and expression were significantly associated with Luminal A type (p = 0.017 and p = 0.011, respectively). However, they did not have any clinical and prognostic values for IBC patients. This result suggested that alterations of PIK3CA pathway contribute to the pathogenesis of specific type of IBC.

Comprehensive ensemble in QSAR prediction for drug discovery.

BACKGROUND: Quantitative structure-activity relationship (QSAR) is a computational modeling method for revealing relationships between structural properties of chemical compounds and biological activities. QSAR modeling is essential for drug discovery, but it has many constraints. Ensemble-based machine learning approaches have been used to overcome constraints and obtain reliable predictions. Ensemble learning builds a set of diversified models and combines them. However, the most prevalent approach random forest and other ensemble approaches in QSAR prediction limit their model diversity to a single subject. RESULTS: The proposed ensemble method consistently outperformed thirteen individual models on 19 bioassay datasets and demonstrated superiority over other ensemble approaches that are limited to a single subject. The comprehensive ensemble method is publicly available at http://data.snu.ac.kr/QSAR/ . CONCLUSIONS: We propose a comprehensive ensemble method that builds multi-subject diversified models and combines them through second-level meta-learning. In addition, we propose an end-to-end neural network-based individual classifier that can automatically extract sequential features from a simplified molecular-input line-entry system (SMILES). The proposed individual models did not show impressive results as a single model, but it was considered the most important predictor when combined, according to the interpretation of the meta-learning.

LncRNAnet: long non-coding RNA identification using deep learning.

Motivation: Long non-coding RNAs (lncRNAs) are important regulatory elements in biological processes. LncRNAs share similar sequence characteristics with messenger RNAs, but they play completely different roles, thus providing novel insights for biological studies. The development of next-generation sequencing has helped in the discovery of lncRNA transcripts. However, the experimental verification of numerous transcriptomes is time consuming and costly. To alleviate these issues, a computational approach is needed to distinguish lncRNAs from the transcriptomes. Results: We present a deep learning-based approach, lncRNAnet, to identify lncRNAs that incorporates recurrent neural networks for RNA sequence modeling and convolutional neural networks for detecting stop codons to obtain an open reading frame indicator. lncRNAnet performed clearly better than the other tools for sequences of short lengths, on which most lncRNAs are distributed. In addition, lncRNAnet successfully learned features and showed 7.83%, 5.76%, 5.30% and 3.78% improvements over the alternatives on a human test set in terms of specificity, accuracy, F1-score and area under the curve, respectively. Availability and implementation: Data and codes are available in http://data.snu.ac.kr/pub/lncRNAnet.

Changes in Lesional and Non-lesional Skin Microbiome During Treatment of Atopic Dermatitis.

The aim of this study was to evaluate changes in the skin surface microbiome in patients with atopic dermatitis during treatment. The effect of narrowband ultraviolet B phototherapy was also studied to determine the influence of exposure to ultraviolet. A total of 18 patients with atopic dermatitis were included in the study. Patients were divided into 2 groups based on treatment: 1 group treated with narrowband ultraviolet B phototherapy and topical corticosteroid, and the other group treated with topical corticosteroid only. Skin swabs and high-throughput sequencing of 16S ribosomal RNA bacterial genes were performed at 3 time-points. The microbial diversity of lesional skin increased greatly after treatment. The proportion of Staphylococcus aureus showed a significant positive correlation with eczema severity. In conclusion, a drastic increase in microbial diversity and decrease in S. aureus proportion were observed with eczema treatment. Narrowband ultraviolet B treatment did not exert additive effects on eczema improvement; however, it appeared to reduce the recurrence of eczema.

Role of Leukotriene B4 Receptor-2 in Mast Cells in Allergic Airway Inflammation.

Mast cells are effector cells in the immune system that play an important role in the allergic airway inflammation. Recently, it was reported that BLT2, a low-affinity leukotriene (LT) B4 receptor, plays a pivotal role in the pathogenesis of allergic airway inflammation through its action in mast cells. We observed that highly elevated expression levels of BLT2 are critical for the pathogenesis leading to allergic airway inflammation, and that if BLT2 expression is downregulated by siBLT2-mediated knockdown, allergic inflammation is dramatically alleviated. Furthermore, we demonstrated that BLT2 mediates the synthesis of vascular endothelial growth factor (VEGF) and Th2 cytokines, such as interleukin (IL)-13, in mast cells during allergic inflammation. Based on the critical roles of BLT2 in mast cells in allergic inflammation, anti-BLT2 strategies could contribute to the development of new therapies for allergic airway inflammation.

TZAP Mutation Leads to Poor Prognosis of Patients with Breast Cancer †.

Background and Objectives: ZBTB48 is a telomere-associated factor that has been renamed as telomeric zinc finger-associated protein (TZAP). It binds preferentially to long telomeres, competing with telomeric repeat factors 1 and 2. Materials and Methods: We analyzed the TZAP mutation in 128 breast carcinomas (BCs). In addition, its association with telomere length was investigated. Results: The TZAP mutation (c.1272 G > A, L424L) was found in 7.8% (10/128) of the BCs and was associated with the N0 stage. BCs with the TZAP mutation had longer telomeres than those without this mutation. Survival analysis showed that the TZAP mutation resulted in poorer overall survival. Conclusions: These results suggest that the TZAP mutation is a possible prognostic marker in BC.

Induction of delayed wound healing by irradiation with optional mechanical compression in swine.

A chronic wound or non-healing wound is one that fails to heal for at least 30 days after injury. This study was designed to create delayed wound healing induced by irradiation and mechanical compression using silicone block. Two female pigs received a single fraction of 20 Gy with 6-MeV electrons to a 22 × 60 cm field on the dorsal body skin 7 weeks before experimentation. A 30 × 30 mm sized wounds were created with preservation of muscle fascia on the dorsum. In groups of six, wounds were designated to be control (C) or test areas of irradiation only (T0), irradiation with silicone blocks for 1 week (T1), irradiation with silicone blocks for 2 weeks (T2), and irradiation with silicone blocks for 3 weeks (T3). Wound contraction, bacterial culture, and histological analysis were performed at 1-week intervals for 4 weeks. Control wounds displayed complete re-epithelialization at Weeks 4; however, all experimental groups (T0, T1, T2, and T3 groups) showed necrosis and delayed healing at Week 4. The number of bacterial strains in control wounds differed significantly from values recorded for all experimental groups from Weeks 1-3 (p < 0.05). However, in comparing the various test wounds (T0, T1, T2, and T3 groups), the numbers of strains did not differ significantly from Weeks 1-4. In the histological analysis, the control wound showed a peak influx of acute and chronic inflammatory cell and diminished inflammation thereafter. However, all experimental groups showed no peak in inflammatory score and prolonged chronic inflammation. In conclusion, radiation exposure alone, which triggers intense inflammation and extensive recruitment of inflammatory cells, proved sufficient to prevent re-epithelialization of skin at 30 days. Insertion of silicone blocks had limited effects on promoting delayed wound healing. Consequently, we now recommend using irradiation alone to simulate delayed wound healing in an experimental setting.

CASPER: context-aware scheme for paired-end reads from high-throughput amplicon sequencing.

Merging the forward and reverse reads from paired-end sequencing is a critical task that can significantly improve the performance of downstream tasks, such as genome assembly and mapping, by providing them with virtually elongated reads. However, due to the inherent limitations of most paired-end sequencers, the chance of observing erroneous bases grows rapidly as the end of a read is approached, which becomes a critical hurdle for accurately merging paired-end reads. Although there exist several sophisticated approaches to this problem, their performance in terms of quality of merging often remains unsatisfactory. To address this issue, here we present a context-aware scheme for paired-end reads (CASPER): a computational method to rapidly and robustly merge overlapping paired-end reads. Being particularly well suited to amplicon sequencing applications, CASPER is thoroughly tested with both simulated and real high-throughput amplicon sequencing data. According to our experimental results, CASPER significantly outperforms existing state-of-the art paired-end merging tools in terms of accuracy and robustness. CASPER also exploits the parallelism in the task of paired-end merging and effectively speeds up by multithreading. CASPER is freely available for academic use at http://best.snu.ac.kr/casper.

Systematic Selection of High-Affinity ssDNA Sequences to Carbon Nanotubes.

Single-walled carbon nanotubes (SWCNTs) have gained significant interest for their potential in biomedicine and nanoelectronics. The functionalization of SWCNTs with single-stranded DNA (ssDNA) enables the precise control of SWCNT alignment and the development of optical and electronic biosensors. This study addresses the current gaps in the field by employing high-throughput systematic selection, enriching high-affinity ssDNA sequences from a vast random library. Specific base compositions and patterns are identified that govern the binding affinity between ssDNA and SWCNTs. Molecular dynamics simulations validate the stability of ssDNA conformations on SWCNTs and reveal the pivotal role of hydrogen bonds in this interaction. Additionally, it is demonstrated that machine learning could accurately distinguish high-affinity ssDNA sequences, providing an accessible model on a dedicated webpage (http://service.k-medai.com/ssdna4cnt). These findings open new avenues for high-affinity ssDNA-SWCNT constructs for stable and sensitive molecular detection across diverse scientific disciplines.

Experimental pig model of clinically relevant wound healing delay by intrinsic factors.

The intrinsic factors associated with delayed wounds are ischaemia, infection, the presence of necrotic tissue and the presence of foreign bodies. This study was designed to create adaptable intrinsic factors induced delayed wound healing model in the pig, similar to clinical conditions. Four pigs were used in this study. Ten 4 × 4 cm fresh full-thickness skin defect wounds were made on the back of each pig. Double-flanged blocks made of silicon, a material that induces foreign body reactions, tissue ischaemia and causes local wound infection, were inserted into half of the wounds (n = 5) and removed at weeks 1, 2 or 3. The other wounds (n = 5) were left open for control purposes. Biopsies were obtained at weeks 1-4 for histological examinations and to evaluate inflammatory cytokines, which included tumour necrosis factor-α (TNF-α), interleukin (IL)-1ß and IL-6. A Periscan PIM 3 Laser was used to measure degrees of ischaemia. A Visitrak Digital(®) device was used to measure the sizes of unhealed wounds. After 3 weeks, unhealed mean wound sizes were 30·72 ± 3·36 and 2·91 ± 1·51 cm(2) , in the experimental and control groups, respectively. Laser Doppler showed limited perfusion at wound margins in wounds containing silicone blocks. Histological findings corresponding to the severity of chronic inflammation were markedly higher and persisted until week 3 in the experimental group. The levels of IL-1 and TNF-α in the third week were statistically significantly higher in the model group. Also, in the fourth week, the level of TNF-α in the model group was significantly higher. Epithelialisation was not complete at the end of the fourth week despite removing silicone blocks in the experimental group. On the basis of macroscopic and histological evaluations and molecular assessments, the described wound model has the characteristics of wound ischaemia and necrosis and persistent inflammation, which are very similar to delayed wounds of patients.

METTL3 regulates alternative splicing of cell cycle-related genes via crosstalk between mRNA m6A modifications and splicing factors.

N6-methyladenosine (m6A) modification in RNA affects various aspects of RNA metabolism and regulates gene expression. This modification is modulated by many regulatory proteins, such as m6A methyltransferases (writers), m6A demethylases (erasers), and m6A-binding proteins (readers). Previous studies have suggested that alterations in m6A regulatory proteins induce genome-wide alternative splicing in many cancer cells. However, the functional effects and molecular mechanisms of m6A-mediated alternative splicing have not been fully elucidated. To understand the consequences of this modification on RNA splicing in cancer cells, we performed RNA sequencing and analyzed alternative splicing patterns in METTL3-knockdown osteosarcoma U2OS cells. We detected 1,803 alternatively spliced genes in METTL3-knockdown cells compared to the controls and found that cell cycle-related genes were enriched in differentially spliced genes. A comparison of the published MeRIP-seq data for METTL14 with our RNA sequencing data revealed that 70-87% of alternatively spliced genes had an m6A peak near 1 kb of alternative splicing sites. Among the 19 RNA-binding proteins enriched in alternative splicing sites, as revealed by motif analysis, expression of SFPQ highly correlated with METTL3 expression in 12,839 TCGA pan-cancer patients. We also found that cell cycle-related genes were enriched in alternatively spliced genes of other cell lines with METTL3 knockdown. Taken together, we suggest that METTL3 regulates m6A-dependent alternative splicing, especially in cell cycle-related genes, by regulating the functions of splicing factors such as SFPQ.

Therapeutic Effects of Amnion-Conjugated Chitosan-Alginate Membranes on Diabetic Wounds in an Induced Diabetic Swine Model: An In Vitro and In Vivo Study.

Background Chitosan (CS) is a well-known antimicrobial dressing material. Moreover, widely used amniotic membranes contain growth factors beneficial for wound healing. Herein, we created a novel amnion-conjugated CS-alginate membrane dressing and tested its wound healing potency in a diabetic swine model. Methods The bovine amniotic powder growth factor contents were evaluated by protein assay, and the powder's wound healing effects were assessed in vitro by HaCaT cell scratch closure. In vivo, two minipigs developed streptozotocin-induced diabetes. Serial serum glucose measurements and intravenous glucose tolerance tests were performed to confirm their diabetic status. Twelve square-shaped wounds created on each pig's back were randomly divided into control ( n = 4), CS ( n = 4), and amnion-CS (AC; n = 4) groups and treated accordingly with different dressings. Wound healing in each group was assessed by measuring wound contraction over time, capturing wound perfusion with indocyanine green (ICG) angiography, and histologically analyzing inflammatory markers. Results Amniotic powder elution promoted HaCaT cell migration in the scratch wound model, suggesting its beneficial in vitro wound healing effects. In vivo, the CS and AC groups showed earlier wound contraction initiation and reepithelialization and earlier wound perfusion improvement by ICG angiography than the control group. Additionally, the wound size of the AC group at week 3 was significantly smaller than those in the control group. There was no significant difference in the numbers of acute and chronic inflammatory cells between the groups. Conclusion The amnion-conjugated CS-alginate membrane, as well as CS dressing alone, could be a favorable dressing option for diabetic wounds.

The feasibility of granzyme B levels using an amperometric immunosensor for lung cancer detection.

Background: Low-dose computed tomography (LDCT) has improved the early detection of lung cancer. However, LDCT scans present several disadvantages, including the abundance of false-positive results, which lead to a high socioeconomic cost, psychological burden, and repeated exposure to radiation. Therefore, the identification of complementary biomarkers is needed to select high-risk individuals for LDCT. Here, we showed that granzyme B testing with the novel immunosensor has diagnostic value for identifying patients with lung cancer. Methods: We enrolled 44 patients with lung cancer and 51 health controls at Pusan National University Yangsan Hospital in Korea between March 2018 and September 2019. The immunosensor analyzed serum granzyme B levels, and their association with lung cancer detection was evaluated with machine learning models. Results: Serum granzyme B levels were assessed in samples from patients with lung cancer and healthy individuals. Granzyme B testing showed 100% sensitivity, 80% specificity, and an area under the curve of 0.938 for lung cancer detection. After combining granzyme B testing with clinical predictors such as age, smoking status, or pack-years, results from the five-fold cross-validation with random forest model improved diagnostic accuracy of 92.1%, with a sensitivity, specificity, and area under the curve of 92.0%, 92.1%, and 0.977, respectively. Conclusions: This feasibility study suggested that granzyme B may be utilized to detect lung cancer.

Freeze-dried bovine amniotic membrane as a cell delivery scaffold in a porcine model of radiation-induced chronic wounds.

BACKGROUND: Locoregional stem cell delivery is very important for increasing the efficiency of cell therapy. Amnisite BA (Amnisite) is a freeze-dried amniotic membrane harvested from bovine placenta. The objective of this study was to investigate the retention of cells of the stromal vascular fraction (SVF) on Amnisite and to determine the effects of cell-loaded Amnisite in a porcine radiation-induced chronic wound model. METHODS: Initially, experiments were conducted to find the most suitable hydration and incubation conditions for the attachment of SVF cells extracted from pig fat to Amnisite. Before seeding, SVFs were labeled with PKH67. The SVF cell-loaded Amnisite (group S), Amnisite only (group A), and polyurethane foam (group C) were applied to treat radiation-induced chronic wounds in a porcine model. Biopsy was performed at 10, 14, and 21 days post-operation for histological analysis. RESULTS: Retaining the SVF on Amnisite required 30 minutes for hydration and 1 hour for incubation. A PKH67 fluorescence study showed that Amnisite successfully delivered the SVF to the wounds. In histological analysis, group S showed increased re-epithelialization and revascularization with decreased inflammation at 10 days post-operation. CONCLUSIONS: SVFs had acceptable adherence on hydrated Amnisite, with successful cell delivery to a radiation-induced chronic wound model.

Meningothelial hamartoma of the scalp.

Meningothelial hamartoma is a benign tumor composed of ectopic meningothelial elements in the dermis and subcutaneous tissue. It mainly occurs in the scalp; however, the incidence is extremely low. The origin of meningothelial hamartoma has not been elucidated; nevertheless, it has been theorized that it derives from ectopic meningothelial rests displaced during embryologic development. It can be diagnosed histologically as proliferation of connective tissue elements and cells arranged in solid nests, resembling vascular tumors. On immunohistochemistry, it stains positively for epithelial membrane antigen and vimentin. At least 17 cases have been reported, verifying the rarity of the lesion. We present the case of a 16-year-old male patient with a soft scalp mass which was thought to be a lipoma, but turned out to be a meningothelial hamartoma on histology.

Development and Validation of a Next-Generation Sequencing-Based Multigene Assay to Predict the Prognosis of Estrogen Receptor-Positive, HER2-Negative Breast Cancer.

PURPOSE: Multigene assays provide useful prognostic information regarding hormone receptor (HR)-positive breast cancer. Next-generation sequencing (NGS)-based platforms have numerous advantages including reproducibility and adaptability in local laboratories. This study aimed to develop and validate an NGS-based multigene assay to predict the distant recurrence risk. EXPERIMENTAL DESIGN: In total, 179 genes including 30 reference genes highly correlated with the 21-gene recurrence score (RS) algorithm were selected from public databases. Targeted RNA-sequencing was performed using 250 and 93 archived breast cancer samples with a known RS in the training and verification sets, respectively, to develop the algorithm and NGS-Prognostic Score (NGS-PS). The assay was validated in 413 independent samples with long-term follow-up data on distant metastasis. RESULTS: In the verification set, the NGS-PS and 21-gene RS displayed 91.4% concurrence (85/93 samples). In the validation cohort of 413 samples, area under the receiver operating characteristic curve plotted using NGS-PS values classified for distant recurrence was 0.76. The best NGS-PS cut-off value predicting distant metastasis was 20. Furthermore, 269 and 144 patients were classified as low- and high-risk patients in accordance with the cut-off. Five- and 10-year estimates of distant metastasis-free survival (DMFS) for low- versus high-risk groups were 97.0% versus 77.8% and 93.2% versus 64.4%, respectively. The age-related HR for distant recurrence without chemotherapy was 9.73 (95% CI, 3.59-26.40) and 3.19 (95% CI, 1.40-7.29) for patients aged ≤50 and >50 years, respectively. CONCLUSIONS: The newly developed and validated NGS-based multigene assay can predict the distant recurrence risk in ER-positive, HER2-negative breast cancer.

End-to-End Representation Learning for Chemical-Chemical Interaction Prediction.

Chemical-chemical interaction (CCI) plays a major role in predicting candidate drugs, toxicities, therapeutic effects, and biological functions. CCI is typically inferred from a variety of information; however, CCI has yet not been predicted using a learning-based approach. In other drug analyses, deep learning has been actively used in recent years. However, in most cases, deep learning has been used only for classification even though it has feature extraction capabilities. Thus, in this paper, we propose an end-to-end representation learning method for CCI, named DeepCCI, which includes feature extraction and a learning-based approach. Our proposed architecture is based on the Siamese network. Hidden representations are extracted from a simplified molecular input line entry system (SMILES), which is a string notation representing the chemical structure using weight-shared convolutional neural networks. Subsequently, L1 element-wise distances between the two extracted hidden representations are measured. The performance of DeepCCI is compared with those of 12 fingerprint-method combinations. The proposed DeepCCI shows the best performance in most of the evaluation metrics used. In addition, DeepCCI was experimentally validated to guarantee the commutative property. The automatically extracted features can alleviate the efforts required for manual feature engineering and improve prediction performance.

Mediatory roles of leukotriene B4 receptors in LPS-induced endotoxic shock.

Sepsis, a systemic inflammatory response syndrome caused by infection, is the most common disease in patients treated in intensive care units. Endotoxic shock, the most critical form of sepsis, is caused by gram-negative bacterial infection. However, the detailed mechanism of endotoxic shock remains unclear. In the present study, we observed that the production of leukotriene B4 (LTB4) and 12(S)-hydroxyeicosatetraenoic acid (HETE), inflammatory lipid mediators acting on LTB4 receptors (BLT1 and BLT2), was significantly upregulated in peritoneal lavage fluid (PF) and serum from an LPS-induced endotoxic shock mouse model. Furthermore, BLT1/2-dependent signaling pathways mediated the expression of IL-17, IL-6, and IL-1ß, key cytokines for the development of endotoxic shock, via NF-κB activation in the LPS-induced endotoxic shock mouse model. Additionally, inhibition of BLT1/2 significantly attenuated inflammation and tissue damage associated with endotoxic shock and enhanced the survival rate of mice with this inflammatory complication. Together, these results suggest that LTB4 receptors play critical mediatory roles in the development of endotoxic shock. Our findings point to LTB4 receptors as potential therapeutic targets for the treatment of endotoxic shock.