Image-based deep learning model using DNA methylation data predicts the origin of cancer of unknown primary.

Cancer of unknown primary (CUP) is a rare type of metastatic cancer in which the origin of the tumor is unknown. Since the treatment strategy for patients with metastatic tumors depends on knowing the primary site, accurate identification of the origin site is important. Here, we developed an image-based deep-learning model that utilizes a vision transformer algorithm for predicting the origin of CUP. Using DNA methylation dataset of 8,233 primary tumors from The Cancer Genome Atlas (TCGA), we categorized 29 cancer types into 18 organ classes and extracted 2,312 differentially methylated CpG sites (DMCs) from non-squamous cancer group and 420 DMCs from squamous cell cancer group. Using these DMCs, we created organ-specific DNA methylation images and used them for model training and testing. Model performance was evaluated using 394 metastatic cancer samples from TCGA (TCGA-meta) and 995 samples (693 primary and 302 metastatic cancers) obtained from 20 independent external studies. We identified that the DNA methylation image reveals a distinct pattern based on the origin of cancer. Our model achieved an overall accuracy of 96.95 % in the TCGA-meta dataset. In the external validation datasets, our classifier achieved overall accuracies of 96.39 % and 94.37 % in primary and metastatic tumors, respectively. Especially, the overall accuracies for both primary and metastatic samples of non-squamous cell cancer were exceptionally high, with 96.79 % and 96.85 %, respectively.

Central neurocytoma exhibits radial glial cell signatures with FGFR3 hypomethylation and overexpression.

We explored the genomic events underlying central neurocytoma (CN), a rare neoplasm of the central nervous system, via multiomics approaches, including whole-exome sequencing, bulk and single-nuclei RNA sequencing, and methylation sequencing. We identified FGFR3 hypomethylation leading to FGFR3 overexpression as a major event in the ontogeny of CN that affects crucial downstream events, such as aberrant PI3K-AKT activity and neuronal development pathways. Furthermore, we found similarities between CN and radial glial cells based on analyses of gene markers and CN tumor cells and postulate that CN tumorigenesis is due to dysregulation of radial glial cell differentiation into neurons. Our data demonstrate the potential role of FGFR3 as one of the leading drivers of tumorigenesis in CN.

Rational design of hybrid sensor arrays combined synergistically with machine learning for rapid response to a hazardous gas leak environment in chemical plants.

Combinations of semiconductor metal oxide (SMO) sensors, electrochemical (EC) sensors, and photoionization detection (PID) sensors were used to discriminate chemical hazards on the basis of machine learning. Sensing data inputs were exploited in the form of either numerical or image data formats, and the classification of chemical hazards with high accuracy was achieved in both cases. Even a small amount of gas sensing or purging data (input for ∼30 s) input can be exploited in machine-learning-based gas discrimination. SMO sensors exhibit high performance even in a single-sensor mode, presumably because of the intrinsic cross-sensitivity of metal oxides, which is otherwise considered a major disadvantage of SMO sensors. EC sensors were enhanced through synergistic integration of sensor combinations with machine learning. For precision detection of multiple target analytes, a minimum number of sensors can be proposed for gas detection/discrimination by combining sensors with dissimilar operating principles. The Type I hybrid sensor combines one SMO sensor, one EC sensor, and one PID sensor and is used to identify NH3 gas mixed with sulfur compounds in simulations of NH3 gas leak accidents in chemical plants. The portable remote sensing module made with a Type I hybrid sensor and LTE module can identify mixed NH3 gas with a detection time of 60 s, demonstrating the potential of the proposed system to quickly respond to hazardous gas leak accidents and prevent additional damage to the environment.

Identification of a Complex Karyotype Signature with Clinical Implications in AML and MDS-EB Using Gene Expression Profiling.

Complex karyotype (CK) is associated with a poor prognosis in both acute myeloid leukemia (AML) and myelodysplastic syndrome with excess blasts (MDS-EB). Transcriptomic analyses have improved our understanding of the disease and risk stratification of myeloid neoplasms; however, CK-specific gene expression signatures have been rarely investigated. In this study, we developed and validated a CK-specific gene expression signature. Differential gene expression analysis between the CK and non-CK groups using data from 348 patients with AML and MDS-EB from four cohorts revealed enrichment of the downregulated genes localized on chromosome 5q or 7q, suggesting that haploinsufficiency due to the deletion of these chromosomes possibly underlies CK pathogenesis. We built a robust transcriptional model for CK prediction using LASSO regression for gene subset selection and validated it using the leave-one-out cross-validation method for fitting the logistic regression model. We established a 10-gene CK signature (CKS) predictive of CK with high predictive accuracy (accuracy 94.22%; AUC 0.977). CKS was significantly associated with shorter overall survival in three independent cohorts, and was comparable to that of previously established risk stratification models for AML. Furthermore, we explored of therapeutic targets among the genes comprising CKS and identified the dysregulated expression of superoxide dismutase 1 (SOD1) gene, which is potentially amenable to SOD1 inhibitors.

An integrative approach for exploring the nature of fibroepithelial neoplasms.

BACKGROUND: Malignant phyllodes tumour (MPT) is a rare breast malignancy with epithelial and mesenchymal features. Currently, there are no appropriate research models or effective targeted therapeutic approaches for MPT. METHODS: We collected fresh frozen tissues from nine patients with MPT and performed whole-exome and RNA sequencing. Additionally, we established patient-derived xenograft (PDX) models from patients with MPT and tested the efficacy of targeting dysregulated pathways in MPT using the PDX model from one MPT. RESULTS: MPT has unique molecular characteristics when compared to breast cancers of epithelial origin and can be classified into two groups. The PDX model derived from one patient with MPT showed that the mouse epithelial component increased during tumour growth. Moreover, targeted inhibition of platelet-derived growth factor receptor (PDGFR) and phosphoinositide 3-kinase (PI3K)/mammalian target of rapamycin (mTOR) by imatinib mesylate and PKI-587 showed in vivo tumour suppression effects. CONCLUSIONS: This study revealed the molecular profiles of MPT that can lead to molecular classification and potential targeted therapy, and suggested that the MPT PDX model can be a useful tool for studying the pathogenesis of fibroepithelial neoplasms and for preclinical drug screening to find new therapeutic strategies for MPT.

PCW-A1001, AI-assisted de novo design approach to design a selective inhibitor for FLT-3(D835Y) in acute myeloid leukemia.

Treating acute myeloid leukemia (AML) by targeting FMS-like tyrosine kinase 3 (FLT-3) is considered an effective treatment strategy. By using AI-assisted hit optimization, we discovered a novel and highly selective compound with desired drug-like properties with which to target the FLT-3 (D835Y) mutant. In the current study, we applied an AI-assisted de novo design approach to identify a novel inhibitor of FLT-3 (D835Y). A recurrent neural network containing long short-term memory cells (LSTM) was implemented to generate potential candidates related to our in-house hit compound (PCW-1001). Approximately 10,416 hits were generated from 20 epochs, and the generated hits were further filtered using various toxicity and synthetic feasibility filters. Based on the docking and free energy ranking, the top compound was selected for synthesis and screening. Of these three compounds, PCW-A1001 proved to be highly selective for the FLT-3 (D835Y) mutant, with an IC50 of 764 nM, whereas the IC50 of FLT-3 WT was 2.54 µM.

Genomic characteristics of invasive mucinous adenocarcinoma of the lung with multiple pulmonary sites of involvement.

Invasive mucinous adenocarcinoma (IMA) of the lung frequently presents with diffuse pneumonic-type features or multifocal lesions, which are regarded as a pattern of intrapulmonary metastases. However, the genomics of multifocal IMAs have not been well studied. We performed whole exome sequencing on samples taken from 2 to 5 regions in seven patients with synchronous multifocal IMAs of the lung (24 regions total). Early initiating driver events, such as KRAS, NKX2-1, TP53, or ARID1A mutations, are clonal mutations and were present in all multifocal IMAs in each patient. The tumor mutational burden of multifocal IMAs was low (mean: 1.13/mega base), but further analyses suggested intra-tumor heterogeneity. The mutational signature analysis found that IMAs were predominantly associated with endogenous mutational process (signature 1), APOBEC activity (signatures 2 and 13), and defective DNA mismatch repair (signature 6), but not related to smoking signature. IMAs synchronously located in the bilateral lower lobes of two patients with background usual interstitial pneumonia had different mutation types, suggesting that they were double primaries. In conclusion, genomic evidence found in this study indicated the clonal intrapulmonary spread of diffuse pneumonic-type or multifocal IMAs, although they can occur in multicentric origins in the background of usual interstitial pneumonia. IMAs exhibited a heterogeneous genomic landscape despite the low somatic mutation burden. Further studies are warranted to determine the clinical significance of the genomic characteristics of IMAs in expanded cohorts.

Integrative analysis of genomic and transcriptomic characteristics associated with progression of aggressive thyroid cancer.

Anaplastic thyroid cancer (ATC) and advanced differentiated thyroid cancers (DTCs) show fatal outcomes, unlike DTCs. Here, we demonstrate mutational landscape of 27 ATCs and 86 advanced DTCs by massively-parallel DNA sequencing, and transcriptome of 13 ATCs and 12 advanced DTCs were profiled by RNA sequencing. TERT, AKT1, PIK3CA, and EIF1AX were frequently co-mutated with driver genes (BRAFV600E and RAS) in advanced DTCs as well as ATC, but tumor suppressors (e.g., TP53 and CDKN2A) were predominantly altered in ATC. CDKN2A loss was significantly associated with poor disease-specific survival in patients with ATC or advanced DTCs, and up-regulation of CD274 (PD-L1) and PDCD1LG2 (PD-L2). Transcriptome analysis revealed a fourth molecular subtype of thyroid cancer (TC), ATC-like, which hardly reflects the molecular signatures in DTC. Furthermore, the activation of JAK-STAT signaling pathway could be a potential druggable target in RAS-positive ATC. Our findings provide insights for precision medicine in patients with advanced TCs.

MicroRNA Expression Profiles in Gastric Carcinogenesis.

Intestinal-type gastric carcinoma exhibits a multistep carcinogenic sequence from adenoma to carcinoma with a gradual increase in genomic alterations. But the roles of microRNAs (miRNA) in this multistage cascade are not fully explored. To identify differentially expressed miRNA (DEM) during early gastric carcinogenesis, we performed miRNA microarray profiling with 24 gastric cancers and precursor lesions (7 early gastric cancer [EGC], 3 adenomas with high-grade dysplasia, 4 adenomas with low-grade dysplasia, and 10 adjacent normal tissues). Alterations in the expression of 132 miRNA were detected; these were categorized into three groups based on their expression patterns. Of these, 42 miRNAs were aberrantly expressed in EGC. Five miRNA (miR-26a, miR-375, miR-574-3p, miR-145, and miR-15b) showed decreased expression since adenoma. Expression of two miRNA, miR-200C and miR-29a, was down-regulated in EGCs compared to normal mucosa or adenomas. Six miRNA (miR-601, miR-107, miR-18a, miR-370, miR-300, and miR-96) showed increased expression in gastric cancer compared to normal or adenoma samples. Five representative miRNAs were further validated with RT-qPCR in independent 77 samples. Taken together, these results suggest that the dysregulated miRNA show alterations at the early stages of gastric tumorigenesis and may be used as a candidate biomarker.

Dysregulated Wnt signalling and recurrent mutations of the tumour suppressor RNF43 in early gastric carcinogenesis.

Several recurrent mutations and epigenetic changes have been identified in advanced gastric cancer, but the genetic alterations associated with early gastric carcinogenesis and malignant transformation remain unclear. We investigated the genomic and transcriptomic landscape of adenomas with low-grade dysplasia (LGD) and high-grade dysplasia (HGD), and intestinal-type early gastric cancer (EGC). The results were validated in an independent cohort that included EGCs directly adjacent to adenoma (EGC-adenomas) that were in the process of malignant transformation, and de novo EGCs that do not seem to have been derived from adenoma. The expression patterns clearly divided into normal, LGD, and EGC, whereas those of HGD overlapped with LGD or EGC. These results suggest that HGD is the critical stage determining malignant transformation. We found that genes related to focal adhesion and extracellular matrix receptor interaction pathways were upregulated as LGD progressed to EGC, whereas canonical Wnt signalling and peroxisome proliferator-activated receptor (PPAR) signalling pathway genes were downregulated in EGC. Genomic alterations such as somatic mutation, gene fusion and copy number variation increased gradually from LGD to EGC. APC mutations were present in 67% of LGDs, 58% of HGDs, and 18% of EGCs. RNF43 mutations were present only in HGD and EGC, and TP53 mutations were present only in EGC. In a validation cohort, RNF43 mutations were present in 35.2% of EGC-adenomas, but in only 8.6% of de novo EGCs. This is the first study to investigate the genomic and transcriptomic landscape of multistep gastric carcinogenesis. We investigated important alterations and their related pathways in each step as tumours progressed from LGD to HGD and eventually to EGC. We suggest that mutations and downregulation of RNF43 may play a critical role in the transition from adenoma to carcinoma. Given these findings and Wnt dependency in tumours with RNF43 mutation, intestinal-type gastric cancer or adenoma with RNF43 mutation might represent a promising indication for Wnt-targeted agents. Copyright © 2016 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.

The full-length DNA sequence of Epstein Barr virus from a human gastric carcinoma cell line, SNU-719.

The consistent presence of Epstein-Barr virus (EBV) in malignant cells of EBV-associated gastric carcinoma (EBVaGC) suggests it plays an important role during the development of EBVaGC. However, the entire genomic sequence of EBV from EBVaGC has yet to be determined. This study first determined, annotated, and analyzed the full genomic sequence of EBV from the naturally infected gastric carcinoma cell line SNU-719 using next-generation sequencing and comparative analyses. In consistent with the notion that EBV sequence isolates better reflect their geographic area than tissue origin, the SNU-719 EBV (named as GC1) was categorized as an East Asian type I EBV. Compared with the prototype B95.8 sequence, SNU-719 EBV contained 1372 variations, with 937 and 435 within coding and non-coding regions, respectively. Of the 937 variations, 465 were non-synonymous changes, while 472 synonymous changes included partial internal deletions in the coding regions of LMP1 and gp350. The RNAseq transcriptome revealed that multiple BART transcripts comprised the majority of EBV RNA reads. The SNU-719 EBV expressed high levels of BART, LF3, BHLF1, and BNLF2. Evidence of RNA editing at multiple sites in the host chromosome was found; however, no evidence of genome integration was seen. The annotated SNU-719 EBV sequence will be a useful reference in future EBVaGC studies.

Genetic association of APOA5 and APOE with metabolic syndrome and their interaction with health-related behavior in Korean men.

BACKGROUND: Genome-wide association studies have been used extensively to identify genetic variants linked to metabolic syndrome (MetS), but most of them have been conducted in non-Asian populations. This study aimed to evaluate the association between MetS and previously studied single nucleotide polymorphisms (SNPs), and their interaction with health-related behavior in Korean men. METHODS: Seventeen SNPs were genotyped and their association with MetS and its components was tested in 1193 men who enrolled in the study at Seoul National University Hospital. RESULTS: We found that rs662799 near APOA5 and rs769450 in APOE had significant association with MetS and its components. The SNP rs662799 was associated with increased risk of MetS, elevated triglyceride (TG) and low levels of high-density lipoprotein, while rs769450 was associated with a decreased risk of TG. The SNPs showed interactions between alcohol drinking and physical activity, and TG levels in Korean men. CONCLUSIONS: We have identified the genetic association and environmental interaction for MetS in Korean men. These results suggest that a strategy of prevention and treatment should be tailored to personal genotype and the population.

Recurrent fusion transcripts detected by whole-transcriptome sequencing of 120 primary breast cancer samples.

Relatively few recurrent gene fusion events have been associated with breast cancer to date. In an effort to uncover novel fusion transcripts, we performed whole-transcriptome sequencing of 120 fresh-frozen primary breast cancer samples and five adjacent normal breast tissues using the Illumina HiSeq2000 platform. Three different fusion-detecting tools (deFuse, Chimerascan, and TopHatFusion) were used, and the results were compared. These tools detected 3,831, 6,630 and 516 fusion transcripts (FTs) overall. We primarily focused on the results obtained using the deFuse software. More FTs were identified from HER2 subtype breast cancer samples than from the luminal or triple-negative subtypes (P < 0.05). Seventy fusion candidates were selected for validation, and 32 (45.7%) were confirmed by RT-PCR and Sanger sequencing. Of the validated fusions, six were recurrent (found in 2 or more samples), three were in-frame (PRDX1-AKR1A1, TACSTD2-OMA1, and C2CD2-TFF1) and three were off-frame (CEACAM7-CEACAM6, CYP4X1-CYP4Z2P, and EEF1DP3-FRY). Notably, the novel read-through fusion, EEF1DP3-FRY, was identified and validated in 6.7% (8/120) of the breast cancer samples. This off-frame fusion results in early truncation of the FRY gene, which plays a key role in the structural integrity during mitosis. Three previously reported fusions, PPP1R1B-STARD3, MFGE8-HAPL, and ETV6-NTRK3, were detected in 8.3, 3.3, and 0.8% of the 120 samples, respectively, by both deFuse and Chimerascan. The recently reported MAGI3-AKT3 fusion was not detected in our analysis. Although future work will be needed to examine the biological significance of our new findings, we identified a number of novel fusions and confirmed some previously reported fusions.

Genetic alterations of JAK/STAT cascade and histone modification in extranodal NK/T-cell lymphoma nasal type.

Extranodal NK/T-cell lymphoma nasal type (ENKL) is a rare type of non-Hodgkin lymphoma that more frequently occurs in East Asia and Latin America. Even though its molecular background has been discussed in the last few years, the current knowledge does not explain the disease pathogenesis in most cases of ENKL. Here, we performed multiple types of next-generation sequencing on 34 ENKL samples, including whole-exome sequencing (9 cancer tissues and 4 cancer cell lines), targeted sequencing (21 cancer tissues), and RNA sequencing (3 cancer tissues and 4 cancer cell lines). Mutations were found most frequently in 3 genes, STAT3, BCOR, and MLL2 (which were present in 9, 7, and 6 cancer samples, respectively), whereas there were only 2 cases of JAK3 mutation. In total, JAK/STAT pathway- and histone modification-related genes accounted for 55.9% and 38.2% of cancer samples, respectively, and their involvement in ENKL pathogenesis was also supported by gene expression analysis. In addition, we provided 177 genes upregulated only in cancer tissues, which appear to be linked with angiocentric and angiodestructive growth of ENKL. In this study, we propose several novel driver genes of ENKL, and show that these genes and their functional groups may be future therapeutic targets of this disease.

Deregulation of immune response genes in patients with Epstein-Barr virus-associated gastric cancer and outcomes.

BACKGROUND & AIMS: Patients with Epstein-Barr virus-associated gastric carcinoma (EBVaGC) have a better prognosis than those with gastric cancer not associated with EBV infection (EBVnGC). This is partly because EBV infection recruits lymphocytes, which infiltrate the tumor. A high degree of tumor heterogeneity is likely to be associated with poor response. We investigated differences in gene expression patterns between EBVaGC and EBVnGC. METHODS: We used gene expression profile analysis to compare tumor and nontumor gastric tissues from 12 patients with EBVaGC and 14 patients with EBVnGC. Findings were validated by whole transcriptome RNAseq and real-time quantitative polymerase chain reaction analyses. CD3(+) primary T cells were isolated from human blood samples; migration of these cells and of Jurkat cells were measured in culture with EBV-infected and uninfected gastric cancer cells. RESULTS: Based on Pearson correlation matrix analysis, EBVaGCs had a higher degree of homogeneity than EBVnGCs. Although 4550 genes were differentially expressed between tumor and nontumor gastric tissues of patients with EBVnGC, only 186 genes were differentially expressed between tumor and nontumor gastric tissues of patients with EBVaGC (P < .001). This finding supports the concept that EBVaGCs have fewer genetic and epigenetic alterations than EBVnGCs. Expression of major histocompatibility complex class II genes and genes that regulate chemokine activity were more often deregulated in EBVaGCs compared with nontumor tissues. In culture, more T cells migrated to EBV-infected gastric cancer cells than to uninfected cells; migration was blocked with a neutralizing antibody against CXCR3 (a receptor for many chemokines). CONCLUSIONS: Fewer genes are deregulated in EBVaGC than in EBVnGC. Most changes in EBVaGCs occur in immune response genes. These changes might allow EBVaGC to recruit reactive immune cells; this might contribute to the better outcomes of these patients compared with those with EBVnGC.

Nonvolatile resistance switching on two-dimensional electron gas.

Two-dimensional electron gas (2DEG) at the complex oxide interfaces have brought about considerable interest for the application of the next-generation multifunctional oxide electronics due to the exotic properties that do not exist in the bulk. In this study, we report the integration of 2DEG into the nonvolatile resistance switching cell as a bottom electrode, where the metal-insulator transition of 2DEG by an external field serves to significantly reduce the OFF-state leakage current while enhancing the on/off ratio. Using the Pt/Ta2O5-y/Ta2O5-x/SrTiO3 heterostructure as a model system, we demonstrate the nonvolatile resistance switching memory cell with a large on/off ratio (>10(6)) and a low leakage current at the OFF state (∼10(-13) A). Beyond exploring nonvolatile memory, our work also provides an excellent framework for exploring the fundamental understanding of novel physics in which electronic and ionic processes are coupled in the complex heterostructures.

Shear Behavior Models of Steel Fiber Reinforced Concrete Beams Modifying Softened Truss Model Approaches.

Recognizing that steel fibers can supplement the brittle tensile characteristics of concrete, many studies have been conducted on the shear performance of steel fiber reinforced concrete (SFRC) members. However, previous studies were mostly focused on the shear strength and proposed empirical shear strength equations based on their experimental results. Thus, this study attempts to estimate the strains and stresses in steel fibers by considering the detailed characteristics of steel fibers in SFRC members, from which more accurate estimation on the shear behavior and strength of SFRC members is possible, and the failure mode of steel fibers can be also identified. Four shear behavior models for SFRC members have been proposed, which have been modified from the softened truss models for reinforced concrete members, and they can estimate the contribution of steel fibers to the total shear strength of the SFRC member. The performances of all the models proposed in this study were also evaluated by a large number of test results. The contribution of steel fibers to the shear strength varied from 5% to 50% according to their amount, and the most optimized volume fraction of steel fibers was estimated as 1%-1.5%, in terms of shear performance.

Metal-induced crystallization of amorphous Si thin films assisted by atomic layer deposition of nickel oxide layers.

Atomic layer deposition (ALD) of nickel oxide was applied to the nickel-induced crystallization of amorphous Si thin films. The nickel-induced crystallization was monitored as a function of annealing temperature and time using Raman spectroscopy. Since Raman spectroscopy allows for the numerical quantification of structural components, the incubation time and the crystallization rates were estimated as functions of the annealing temperature. The spatial locations of a nickel-based species, probably NiSi2, were investigated using X-ray photoelectron spectrometry. The formed NiSi2 seeds appeared to accelerate the crystallization kinetics in amorphous Si thin films deposited onto glass substrates. The ramifications of the atomic layer deposition are discussed with regard to large-panel displays, with special emphasis on the sophisticated control of the catalytic elements, especially nickel.