Low-rank regression models for multiple binary responses and their applications to cancer cell-line encyclopedia data.

In this paper, we study high-dimensional multivariate logistic regression models in which a common set of covariates is used to predict multiple binary outcomes simultaneously. Our work is primarily motivated from many biomedical studies with correlated multiple responses such as the cancer cell-line encyclopedia project. We assume that the underlying regression coefficient matrix is simultaneously low-rank and row-wise sparse. We propose an intuitively appealing selection and estimation framework based on marginal model likelihood, and we develop an efficient computational algorithm for inference. We establish a novel high-dimensional theory for this nonlinear multivariate regression. Our theory is general, allowing for potential correlations between the binary responses. We propose a new type of nuclear norm penalty using the smooth clipped absolute deviation, filling the gap in the related non-convex penalization literature. We theoretically demonstrate that the proposed approach improves estimation accuracy by considering multiple responses jointly through the proposed estimator when the underlying coefficient matrix is low-rank and row-wise sparse. In particular, we establish the non-asymptotic error bounds, and both rank and row support consistency of the proposed method. Moreover, we develop a consistent rule to simultaneously select the rank and row dimension of the coefficient matrix. Furthermore, we extend the proposed methods and theory to a joint Ising model, which accounts for the dependence relationships. In our analysis of both simulated data and the cancer cell line encyclopedia data, the proposed methods outperform the existing methods in better predicting responses.

Debiased inference for heterogeneous subpopulations in a high-dimensional logistic regression model.

Due to the prevalence of complex data, data heterogeneity is often observed in contemporary scientific studies and various applications. Motivated by studies on cancer cell lines, we consider the analysis of heterogeneous subpopulations with binary responses and high-dimensional covariates. In many practical scenarios, it is common to use a single regression model for the entire data set. To do this effectively, it is critical to quantify the heterogeneity of the effect of covariates across subpopulations through appropriate statistical inference. However, the high dimensionality and discrete nature of the data can lead to challenges in inference. Therefore, we propose a novel statistical inference method for a high-dimensional logistic regression model that accounts for heterogeneous subpopulations. Our primary goal is to investigate heterogeneity across subpopulations by testing the equivalence of the effect of a covariate and the significance of the overall effects of a covariate. To achieve overall sparsity of the coefficients and their fusions across subpopulations, we employ a fused group Lasso penalization method. In addition, we develop a statistical inference method that incorporates bias correction of the proposed penalized method. To address computational issues due to the nonlinear log-likelihood and the fused Lasso penalty, we propose a computationally efficient and fast algorithm by adapting the ideas of the proximal gradient method and the alternating direction method of multipliers (ADMM) to our settings. Furthermore, we develop non-asymptotic analyses for the proposed fused group Lasso and prove that the debiased test statistics admit chi-squared approximations even in the presence of high-dimensional variables. In simulations, the proposed test outperforms existing methods. The practical effectiveness of the proposed method is demonstrated by analyzing data from the Cancer Cell Line Encyclopedia (CCLE).

Quantile forward regression for high-dimensional survival data.

Despite the urgent need for an effective prediction model tailored to individual interests, existing models have mainly been developed for the mean outcome, targeting average people. Additionally, the direction and magnitude of covariates' effects on the mean outcome may not hold across different quantiles of the outcome distribution. To accommodate the heterogeneous characteristics of covariates and provide a flexible risk model, we propose a quantile forward regression model for high-dimensional survival data. Our method selects variables by maximizing the likelihood of the asymmetric Laplace distribution (ALD) and derives the final model based on the extended Bayesian Information Criterion (EBIC). We demonstrate that the proposed method enjoys a sure screening property and selection consistency. We apply it to the national health survey dataset to show the advantages of a quantile-specific prediction model. Finally, we discuss potential extensions of our approach, including the nonlinear model and the globally concerned quantile regression coefficients model.

The role of myokines in cancer: crosstalk between skeletal muscle and tumor.

Loss of skeletal muscle mass is a primary feature of sarcopenia and cancer cachexia. In cancer patients, tumor-derived inflammatory factors promote muscle atrophy via tumor-to-muscle effects, which is closely associated with poor prognosis. During the past decade, skeletal muscle has been considered to function as an autocrine, paracrine, and endocrine organ by releasing numerous myokines. The circulating myokines can modulate pathophysiology in the other organs, as well as in the tumor microenvironment, suggesting myokines function as muscleto-tumor signaling molecules. Here, we highlight the roles of myokines in tumorigenesis, particularly in terms of crosstalk between skeletal muscle and tumor. Better understanding of tumor-to-muscle and muscle-to-tumor effects will shed light on novel strategies for the diagnosis and treatment of cancer. [BMB Reports 2023; 56(7): 365-373].

Varying-coefficients for regional quantile via KNN-based LASSO with applications to health outcome study.

Health outcomes, such as body mass index and cholesterol levels, are known to be dependent on age and exhibit varying effects with their associated risk factors. In this paper, we propose a novel framework for dynamic modeling of the associations between health outcomes and risk factors using varying-coefficients (VC) regional quantile regression via K-nearest neighbors (KNN) fused Lasso, which captures the time-varying effects of age. The proposed method has strong theoretical properties, including a tight estimation error bound and the ability to detect exact clustered patterns under certain regularity conditions. To efficiently solve the resulting optimization problem, we develop an alternating direction method of multipliers (ADMM) algorithm. Our empirical results demonstrate the efficacy of the proposed method in capturing the complex age-dependent associations between health outcomes and their risk factors.

SCADIE: simultaneous estimation of cell type proportions and cell type-specific gene expressions using SCAD-based iterative estimating procedure.

A challenge in bulk gene differential expression analysis is to differentiate changes due to cell type-specific gene expression and cell type proportions. SCADIE is an iterative algorithm that simultaneously estimates cell type-specific gene expression profiles and cell type proportions, and performs cell type-specific differential expression analysis at the group level. Through its unique penalty and objective function, SCADIE more accurately identifies cell type-specific differentially expressed genes than existing methods, including those that may be missed from single cell RNA-Seq data. SCADIE has robust performance with respect to the choice of deconvolution methods and the sources and quality of input data.

Direct Contact with Platelets Induces Podoplanin Expression and Invasion in Human Oral Squamous Cell Carcinoma Cells.

Oral squamous cell carcinoma (OSCC) is mostly diagnosed at an advanced stage, with local and/or distal metastasis. Thus, locoregional and/or local control of the primary tumor is crucial for a better prognosis in patients with OSCC. Platelets have long been considered major players in cancer metastasis. Traditional antiplatelet agents, such as aspirin, are thought to be potential chemotherapeutics, but they need to be used with caution because of the increased bleeding risk. Podoplanin (PDPN)-expressing cancer cells can activate platelets and promote OSCC metastasis. However, the reciprocal effect of platelets on PDPN expression in OSCC has not been investigated. In this study, we found that direct contact with platelets upregulated PDPN and integrin ß1 at the protein level and promoted invasiveness of human OSCC Ca9.22 cells that express low levels of PDPN. In another human OSCC HSC3 cell line that express PDPN at an abundant level, silencing of the PDPN gene reduced cell invasiveness. Analysis of the public database further supported the co-expression of PDPN and integrin ß1 and their increased expression in metastatic tissues compared to normal and tumor tissues of the oral cavity. Taken together, these data suggest that PDPN is a potential target to regulate platelet-tumor interaction and metastasis for OSCC treatment, which can overcome the limitations of traditional antiplatelet drugs.

A Transformer-Based Neural Machine Translation Model for Arabic Dialects That Utilizes Subword Units.

Languages that allow free word order, such as Arabic dialects, are of significant difficulty for neural machine translation (NMT) because of many scarce words and the inefficiency of NMT systems to translate these words. Unknown Word (UNK) tokens represent the out-of-vocabulary words for the reason that NMT systems run with vocabulary that has fixed size. Scarce words are encoded completely as sequences of subword pieces employing the Word-Piece Model. This research paper introduces the first Transformer-based neural machine translation model for Arabic vernaculars that employs subword units. The proposed solution is based on the Transformer model that has been presented lately. The use of subword units and shared vocabulary within the Arabic dialect (the source language) and modern standard Arabic (the target language) enhances the behavior of the multi-head attention sublayers for the encoder by obtaining the overall dependencies between words of input sentence for Arabic vernacular. Experiments are carried out from Levantine Arabic vernacular (LEV) to modern standard Arabic (MSA) and Maghrebi Arabic vernacular (MAG) to MSA, Gulf-MSA, Nile-MSA, Iraqi Arabic (IRQ) to MSA translation tasks. Extensive experiments confirm that the suggested model adequately addresses the unknown word issue and boosts the quality of translation from Arabic vernaculars to Modern standard Arabic (MSA).

Oral-Gut Microbiome Axis in Gastrointestinal Disease and Cancer.

It is well-known that microbiota dysbiosis is closely associated with numerous diseases in the human body. The oral cavity and gut are the two largest microbial habitats, playing a major role in microbiome-associated diseases. Even though the oral cavity and gut are continuous regions connected through the gastrointestinal tract, the oral and gut microbiome profiles are well-segregated due to the oral-gut barrier. However, the oral microbiota can translocate to the intestinal mucosa in conditions of the oral-gut barrier dysfunction. Inversely, the gut-to-oral microbial transmission occurs as well in inter- and intrapersonal manners. Recently, it has been reported that oral and gut microbiomes interdependently regulate physiological functions and pathological processes. Oral-to-gut and gut-to-oral microbial transmissions can shape and/or reshape the microbial ecosystem in both habitats, eventually modulating pathogenesis of disease. However, the oral-gut microbial interaction in pathogenesis has been underappreciated to date. Here, we will highlight the oral-gut microbiome crosstalk and its implications in the pathogenesis of the gastrointestinal disease and cancer. Better understanding the role of the oral-gut microbiome axis in pathogenesis will be advantageous for precise diagnosis/prognosis and effective treatment.

Platelet CLEC2-Podoplanin Axis as a Promising Target for Oral Cancer Treatment.

Cancer tissues are not just simple masses of malignant cells, but rather complex and heterogeneous collections of cellular and even non-cellular components, such as endothelial cells, stromal cells, immune cells, and collagens, referred to as tumor microenvironment (TME). These multiple players in the TME develop dynamic interactions with each other, which determines the characteristics of the tumor. Platelets are the smallest cells in the bloodstream and primarily regulate blood coagulation and hemostasis. Notably, cancer patients often show thrombocytosis, a status of an increased platelet number in the bloodstream, as well as the platelet infiltration into the tumor stroma, which contributes to cancer promotion and progression. Thus, platelets function as one of the important stromal components in the TME, emerging as a promising chemotherapeutic target. However, the use of traditional antiplatelet agents, such as aspirin, has limitations mainly due to increased bleeding complications. This requires to implement new strategies to target platelets for anti-cancer effects. In oral squamous cell carcinoma (OSCC) patients, both high platelet counts and low tumor-stromal ratio (high stroma) are strongly correlated with increased metastasis and poor prognosis. OSCC tends to invade adjacent tissues and bones and spread to the lymph nodes for distant metastasis, which is a huge hurdle for OSCC treatment in spite of relatively easy access for visual examination of precancerous lesions in the oral cavity. Therefore, locoregional control of the primary tumor is crucial for OSCC treatment. Similar to thrombocytosis, higher expression of podoplanin (PDPN) has been suggested as a predictive marker for higher frequency of lymph node metastasis of OSCC. Cumulative evidence supports that platelets can directly interact with PDPN-expressing cancer cells via C-type lectin-like receptor 2 (CLEC2), contributing to cancer cell invasion and metastasis. Thus, the platelet CLEC2-PDPN axis could be a pinpoint target to inhibit interaction between platelets and OSCC, avoiding undesirable side effects. Here, we will review the role of platelets in cancer, particularly focusing on CLEC2-PDPN interaction, and will assess their potentials as therapeutic targets for OSCC treatment.

Integrating multidimensional data for clustering analysis with applications to cancer patient data.

Advances in high-throughput genomic technologies coupled with large-scale studies including The Cancer Genome Atlas (TCGA) project have generated rich resources of diverse types of omics data to better understand cancer etiology and treatment responses. Clustering patients into subtypes with similar disease etiologies and/or treatment responses using multiple omics data types has the potential to improve the precision of clustering than using a single data type. However, in practice, patient clustering is still mostly based on a single type of omics data or ad hoc integration of clustering results from individual data types, leading to potential loss of information. By treating each omics data type as a different informative representation from patients, we propose a novel multi-view spectral clustering framework to integrate different omics data types measured from the same subject. We learn the weight of each data type as well as a similarity measure between patients via a non-convex optimization framework. We solve the proposed non-convex problem iteratively using the ADMM algorithm and show the convergence of the algorithm. The accuracy and robustness of the proposed clustering method is studied both in theory and through various synthetic data. When our method is applied to the TCGA data, the patient clusters inferred by our method show more significant differences in survival times between clusters than those inferred from existing clustering methods.

Randomized Controlled Trial of the Caregiver Orientation for Mobilizing Personal Assets and Strengths for Self-Care (COMPASS) for Caregiving Journey: A National Family Caregiver Support Program in a Long-Term Care Insurance System.

OBJECTIVES: To investigate the effects of a national support program on family caregivers for long-term care (LTC) recipients. DESIGN: A single-blinded randomized controlled trial compared the 8-week Caregiver Orientation for Mobilizing Personal Assets and Strengths for Self-Care (COMPASS) program consisting of 6 individual in-home, 3 group support, and 2 telephone sessions with a multicomponent intervention, and a control group. SETTING AND PARTICIPANTS: In total, 969 caregivers who were living with LTC recipients assessed as having a high caregiving burden in 12 Korean cities. MEASURES: The primary outcomes were depression, burden, and stress levels of caregivers, the secondary outcomes were caregiver self-efficacy, positive aspects of caregiving, social support, social activities, and health risk behaviors. These outcomes were measured at baseline and after the 8-week program, analyzed using modified intention-to-treat, per-protocol (PP), and non-PP analyses. RESULTS: The modified intention-to-treat analysis revealed significant improvements in burden (effect size, = 0.010, P = .008), depression (ηp2 = 0.012, P = .003), and health risk behaviors (ηp2 = 0.010, P = .012) for the experimental group compared with the control group. However, there were no significant differences between the 2 groups in improving stress (P = .997), social support (P = .234), or social activities (P = .816). The PP analysis indicated that the COMPASS program was successful in increasing positive aspects of caregiving (ηp2 = 0.013, P = .004) and self-efficacy (ηp2 = 0.010, P = .032) compared with the control group. CONCLUSIONS AND IMPLICATIONS: The COMPASS program was effective in family caregivers of LTC recipients in critical aspects of physical and psychological outcomes, especially in demonstrating the important role of participating in group support sessions. It is feasible for the program to become a formal national support program as part of the national insurance system in Republic of Korea.

Changes in Emberiza bunting communities and populations spanning 100 years in Korea.

The rapid decline of a few Emberiza bunting species is increasing conservation concerns, especially in Asia. However, temporal changes in communities and populations of buntings, ones of the most common migratory songbirds in Korea, have not been quantitatively assessed. To understand how the status of buntings has changed over the past 100 years, we collated abundance data from museum collections and bird-banding records between 1910 and 2019. We also used presence-absence data for buntings collected by a nationwide census scheme between 1997 to 2012. Our analysis showed that bunting communities reconstructed from museum-specimen and bird-banding data were not significantly different; however, community composition differed over time. The Meadow (E. cioides), Yellow-throated (E. elegans), Black-faced (E. spodocephala), Rustic (E. rustica) and Chestnut Buntings (E. rutila), which are still common or were once common species, significantly affected the temporal changes in bunting community composition. There were no recent changes in the presence of Rustic and Chestnut Buntings since 1997, but they caused medium-term changes in the bunting community composition, suggesting that there was a sharp to moderate decline in their numbers in the past. The probability of the presence of six bunting species decreased annually, with the most prominent decline in two common breeders, the Meadow (-2.99%/year) and Yellow-throated Buntings (-1.82%/year). This finding suggests that breeding buntings in Korea are under high pressure, as are the migratory buntings. Moreover, despite its recent population decline, the Yellow-throated Bunting was still a major contributor to the community, suggesting that bunting diversity has also been deteriorating while bunting populations are shrinking. Long-term monitoring schemes across their distribution ranges, international cooperation for identifying major threats and key areas of conservation, and law enforcement against illegal hunting and habitat loss are strongly required to mitigate the on-going decline of buntings in Korea and Asia.

STAT3 Stabilizes IKKα Protein through Direct Interaction in Transformed and Cancerous Human Breast Epithelial Cells.

Signal transducer and activator of transcription 3 (STAT3) and nuclear factor-κB (NF-κB) are two representative transcription factors that play a critical role in inflammation-associated tumorigenesis through multi-level cooperation. Unlike other types of tumors, breast carcinomas have shown a significant dependency on the non-classical NF-κB pathway as well as the classical one. The α subunit of the inhibitor of the κB kinase (IKK) complex, IKKα, is involved in both classical and non-classical activation of NF-κB. Although the cross-talk between STAT3 and NF-κB has been suggested in several studies, the interplay between STAT3 and the regulators of NF-κB including IKKα has not been fully clarified yet. In this study, we observed overexpression and co-localization of IKKα and STAT3 in human breast cancer tissues as well as in H-Ras transformed human breast epithelial (H-Ras MCF-10A) and breast cancer (MDA-MB-231) cells. By utilizing small interfering RNA (siRNA) technology, we were able to demonstrate that STAT3 up-regulated IKKα, but not IKKß or IKKγ, in these cells. This was attributable to direct binding to and subsequent stabilization of IKKα protein by blocking the ubiquitin-proteasome system. Notably, we identified the lysine 44 residue of IKKα as a putative binding site for STAT3. Moreover, siRNA knockdown of IKKα attenuated viability, anchorage-independent growth and migratory capabilities of H-Ras MCF-10A cells. Taken together, these findings propose a novel mechanism responsible for NF-κB activation by STAT3 through stabilization of IKKα, which contributes to breast cancer promotion and progression. Thus, breaking the STAT3-IKKα alliance can be an alternative therapeutic strategy for the treatment of breast cancer.

Production of bio-oil with reduced polycyclic aromatic hydrocarbons via continuous pyrolysis of biobutanol process derived waste lignin.

The fast pyrolysis of waste lignin derived from biobutanol production process was performed to determine the optimal pyrolysis conditions and pyrolysis product properties. Four types of pyrolysis reactors, e.g.: micro-scale pyrolyzer-gas chromatography/mass spectrometry, lab and bench scale fixed bed (FB) reactors, and bench scale rotary kiln (RK) reactor, were employed to compare the pyrolysis reaction conditions and product properties obtained from different reactors. The yields of char, oil, and gas obtained from lab scale and bench scale reactor were almost similar compared to FB reactor. RK reactor produced desirable bio-oil with much reduced yield of poly aromatic hydrocarbons (cancer precursor) due to its higher cracking reaction efficiency. In addition, char agglomeration and foaming of lignin pyrolysis were greatly restricted by using RK reactor compared to the FB reactor.

NITUMID: Nonnegative matrix factorization-based Immune-TUmor MIcroenvironment Deconvolution.

MOTIVATION: A number of computational methods have been proposed recently to profile tumor microenvironment (TME) from bulk RNA data, and they have proved useful for understanding microenvironment differences among therapeutic response groups. However, these methods are not able to account for tumor proportion nor variable mRNA levels across cell types. RESULTS: In this article, we propose a Nonnegative Matrix Factorization-based Immune-TUmor MIcroenvironment Deconvolution (NITUMID) framework for TME profiling that addresses these limitations. It is designed to provide robust estimates of tumor and immune cells proportions simultaneously, while accommodating mRNA level differences across cell types. Through comprehensive simulations and real data analyses, we demonstrate that NITUMID not only can accurately estimate tumor fractions and cell types' mRNA levels, which are currently unavailable in other methods; it also outperforms most existing deconvolution methods in regular cell type profiling accuracy. Moreover, we show that NITUMID can more effectively detect clinical and prognostic signals from gene expression profiles in tumor than other methods. AVAILABILITY AND IMPLEMENTATION: The algorithm is implemented in R. The source code can be downloaded at https://github.com/tdw1221/NITUMID. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Estimation of Health Utilities Based on the Response to Treatment in Atopic Dermatitis: a Population-based Study.

PURPOSE: This study estimated utility weights based on the response to treatment for atopic dermatitis in the general population. METHODS: The Korean general population aged 20-60 years was stratified by using a random sampling method based on age and sex. Two hypothetical health states of atopic dermatitis were developed: response to treatment and no response to treatment. Health utility values were estimated by using time trade-off (TTO) based on a period of 10 years, TTO based on life expectancy, and EuroQol 5-Dimension (EQ-5D) including a visual analog scale (VAS). The mean utility value and 95% CI were derived, and comparisons of subgroups using the t test and ANOVA were performed. We conducted a multilevel analysis after controlling the sociodemographic variables to consider repeated measures. FINDINGS: A total of 155 participants were included in the survey. Their mean age was 39.7 years; 58.7% of participants were women. The mean health utility values for response and no response using TTO based on 10 years were 0.847 and 0.380, respectively. The estimated health utility values of response and no response were 0.865 and 0.476 using TTO based on life expectancy, and 0.814 and 0.279 using EQ-5D. For VAS, the response and no response were 0.744 and 0.322. After controlling the covariates, the important factors that affected utility values were response and no response to treatment (P < 0.001). IMPLICATIONS: This study showed that the utility weights of people with no response to atopic dermatitis treatment were lower compared with response from the general population. Health care providers should therefore consider symptom control as an important factor affecting the quality of life of those with atopic dermatitis.

Design of thin-film photonic crystals with complete photonic bandgap.

We theoretically investigate the optical characteristics of a thin-film photonic crystal structure with a complete photonic bandgap for both polarization of the transverse electric and transverse magnetic modes for any in-plane direction. The structure consists of three-layer stacked two-dimensional photonic crystal slabs, and the thickness of the structure is less than a few wavelengths. We show that a wide complete photonic bandgap can be obtained in the asymmetrically stacked photonic crystal structure. In addition, we designed a waveguide with a broad bandwidth of 100 nm and a nanocavity with a quality factor of 3.7 × 107 in the structures.

Spectral clustering based on learning similarity matrix.

Motivation: Single-cell RNA-sequencing (scRNA-seq) technology can generate genome-wide expression data at the single-cell levels. One important objective in scRNA-seq analysis is to cluster cells where each cluster consists of cells belonging to the same cell type based on gene expression patterns. Results: We introduce a novel spectral clustering framework that imposes sparse structures on a target matrix. Specifically, we utilize multiple doubly stochastic similarity matrices to learn a similarity matrix, motivated by the observation that each similarity matrix can be a different informative representation of the data. We impose a sparse structure on the target matrix followed by shrinking pairwise differences of the rows in the target matrix, motivated by the fact that the target matrix should have these structures in the ideal case. We solve the proposed non-convex problem iteratively using the ADMM algorithm and show the convergence of the algorithm. We evaluate the performance of the proposed clustering method on various simulated as well as real scRNA-seq data, and show that it can identify clusters accurately and robustly. Availability and implementation: The algorithm is implemented in MATLAB. The source code can be downloaded at https://github.com/ishspsy/project/tree/master/MPSSC. Supplementary information: Supplementary data are available at Bioinformatics online.

A Neural Machine Translation Model for Arabic Dialects That Utilises Multitask Learning (MTL).

In this research article, we study the problem of employing a neural machine translation model to translate Arabic dialects to modern standard Arabic. The proposed solution of the neural machine translation model is prompted by the recurrent neural network-based encoder-decoder neural machine translation model that has been proposed recently, which generalizes machine translation as sequence learning problems. We propose the development of a multiytask learning (MTL) model which shares one decoder among language pairs, and every source language has a separate encoder. The proposed model can be applied to limited volumes of data as well as extensive amounts of data. Experiments carried out have shown that the proposed MTL model can ensure a higher quality of translation when compared to the individually learned model.