Biomarker identification by interpretable maximum mean discrepancy.

MOTIVATION: In many biomedical applications, we are confronted with paired groups of samples, such as treated versus control. The aim is to detect discriminating features, i.e. biomarkers, based on high-dimensional (omics-) data. This problem can be phrased more generally as a two-sample problem requiring statistical significance testing to establish differences, and interpretations to identify distinguishing features. The multivariate maximum mean discrepancy (MMD) test quantifies group-level differences, whereas statistically significantly associated features are usually found by univariate feature selection. Currently, few general-purpose methods simultaneously perform multivariate feature selection and two-sample testing. RESULTS: We introduce a sparse, interpretable, and optimized MMD test (SpInOpt-MMD) that enables two-sample testing and feature selection in the same experiment. SpInOpt-MMD is a versatile method and we demonstrate its application to a variety of synthetic and real-world data types including images, gene expression measurements, and text data. SpInOpt-MMD is effective in identifying relevant features in small sample sizes and outperforms other feature selection methods such as SHapley Additive exPlanations and univariate association analysis in several experiments. AVAILABILITY AND IMPLEMENTATION: The code and links to our public data are available at https://github.com/BorgwardtLab/spinoptmmd.

Leritrelvir for the treatment of mild or moderate COVID-19 without co-administered ritonavir: a multicentre randomised, double-blind, placebo-controlled phase 3 trial.

Background: Leritrelvir is a novel α-ketoamide based peptidomimetic inhibitor of SARS-CoV-2 main protease. A preclinical study has demonstrated leritrelvir poses similar antiviral activities towards different SARS-CoV-2 variants compared with nirmatrelvir. A phase 2 clinical trial has shown a comparable antiviral efficacy and safety between leritrelvir with and without ritonavir co-administration. This trial aims to test efficacy and safety of leritrelvir monotherapy in adults with mild-to-moderate COVID-19. Methods: This was a randomised, double-blind, placebo-controlled, multicentre phase 3 trial at 29 clinical sites in China. Enrolled patients were from 18 to 75 years old, diagnosed with mild or moderate COVID-19 and not requiring hospitalization. Patients had a positive SARS-CoV-2 nucleic acid test (NAT) and at least one of the COVID-19 symptoms within 48 h before randomization, and the interval between the first positive SARS-CoV-2 NAT and randomization was ≤120 h (5 days). Patients were randomly assigned in a 1:1 ratio to receive a 5-day course of either oral leritrelvir 400 mg TID or placebo. The primary efficacy endpoint was the time from the first dose to sustained clinical recovery of all 11 symptoms (stuffy or runny nose, sore throat, shortness of breath or dyspnea, cough, muscle or body aches, headache, chills, fever ≥37 °C, nausea, vomiting, and diarrhea). The safety endpoint was the incidence of adverse events (AE). Primary and safety analyses were performed in the intention-to-treat (ITT) population. This study is registered with ClinicalTrials.gov, NCT05620160. Findings: Between Nov 12 and Dec 30, 2022 when the zero COVID policy was abolished nationwide, a total of 1359 patients underwent randomization, 680 were assigned to leritrelvir group and 679 to placebo group. The median time to sustained clinical recovery in leritrelvir group was significantly shorter (251.02 h [IQR 188.95-428.68 h]) than that of Placebo (271.33 h [IQR 219.00-529.63 h], P = 0.0022, hazard ratio [HR] 1.20, 95% confidence interval [CI], 1.07-1.35). Further analysis of subgroups for the median time to sustained clinical recovery revealed that (1) subgroup with positive viral nucleic acid tested ≤72 h had a 33.9 h difference in leritrelvir group than that of placebo; (2) the subgroup with baseline viral load >8 log 10 Copies/mL in leritrelvir group had 51.3 h difference than that of placebo. Leritrelvir reduced viral load by 0.82 log10 on day 4 compared to placebo. No participants in either group progressed to severe COVID-19 by day 29. Adverse events were reported in two groups: leritrelvir 315 (46.46%) compared with placebo 292 (43.52%). Treatment-relevant AEs were similar 218 (32.15%) in the leritrelvir group and 186 (27.72%) in placebo. Two cases of COVID-19 pneumonia were reported in placebo group, and one case in leritrelvir group, none of them were considered by the investigators to be leritrelvir related. The most frequently reported AEs (occurring in ≥5% of participants in at least one group) were laboratory finding: hypertriglyceridemia (leritrelvir 79 [11.7%] vs. placebo 70 [10.4%]) and hyperlipidemia (60 [8.8%] vs. 52 [7.7%]); all of them were nonserious. Interpretation: Leritrelvir monotherapy has good efficacy for mild-to-moderate COVID-19 and without serious safety concerns. Funding: This study was funded by the National Multidisciplinary Innovation Team Project of Traditional Chinese Medicine, Guangdong Science and Technology Foundation, Guangzhou Science and Technology Planning Project and R&D Program of Guangzhou Laboratory.

Is Lymphadenectomy Reasonable for Elderly Intrahepatic Cholangiocarcinoma Patients?

BACKGROUND: In this study, we aimed to determine the impact of lymphadenectomy (LND) on clinical outcomes in ICC patients aged ≥ 70 years. METHODS: Four hundred and three eligible patients diagnosed with ICC who underwent hepatectomy between 2004 and 2019 were enrolled in the Surveillance, Epidemiology, and End Results database. The impact of LND on perioperative mortality and overall survival (OS) as well as the optimal total number of lymph nodes examined (TNLE) was estimated. RESULTS: One hundred thirty-nine pairs of patients were matched by propensity score matching. Perioperative mortality was comparable between the LND and non-LND (nLND) groups (0.7% vs. 2.9%, P = 0.367). The median OS in the LND group was significantly longer (44 vs. 32 months, P = 0.045) and LND was identified as an independent protective factor for OS by multivariate analysis (HR 0.65, 95% CI 0.46-0.92, P = 0.014). Patients with the following characteristics were potential beneficiaries of LND: white, female, no/moderate fibrosis, tumor size > 5 cm, solitary tumor, and localized invasion (all P < 0.05). TNLE ≥ 6 had the greatest discriminatory power for identifying lymph node metastasis (area under the curve, 0.704, Youden index, 0.365, P = 0.002). Patients with pathologically confirmed lymph node metastasis are likely to benefit from adjuvant therapy (40 months vs. 4 months, P = 0.052). CONCLUSIONS: Advanced age (≥ 70 years) was not a contraindication for LND, which facilitates accurate nodal staging and guides postoperative management. Appropriately selected elderly populations could benefit from LND.

Cellular and molecular atlas of the placenta from a COVID-19 pregnant woman infected at midgestation highlights the defective impacts on foetal health.

OBJECTIVES: The impacts of the current COVID-19 pandemic on maternal and foetal health are enormous and of serious concern. However, the influence of SARS-CoV-2 infection at early-to-mid gestation on maternal and foetal health remains unclear. MATERIALS AND METHODS: Here, we report the follow-up study of a pregnant woman of her whole infective course of SARS-CoV-2, from asymptomatic infection at gestational week 20 to mild and then severe illness state, and finally cured at Week 24. Following caesarean section due to incomplete uterine rupture at Week 28, histological examinations on the placenta and foetal tissues as well as single-cell RNA sequencing (scRNA-seq) for the placenta were performed. RESULTS: Compared with the gestational age-matched control placentas, the placenta from this COVID-19 case exhibited more syncytial knots and lowered expression of syncytiotrophoblast-related genes. The scRNA-seq analysis demonstrated impaired trophoblast differentiation, activation of antiviral and inflammatory CD8 T cells, as well as the tight association of increased inflammatory responses in the placenta with complement over-activation in macrophages. In addition, levels of several inflammatory factors increased in the placenta and foetal blood. CONCLUSION: These findings illustrate a systematic cellular and molecular signature of placental insufficiency and immune activation at the maternal-foetal interface that may be attributed to SARS-CoV-2 infection at the midgestation stage, which highly suggests the extensive care for maternal and foetal outcomes in pregnant women suffering from COVID-19.

Case Report: Massive Hepatocellular Carcinoma Complete Surgical Resection After Portal Vein Embolization and Multimodality Therapy.

A high proportion of massive patients with hepatocellular carcinoma (HCC) are not amenable for surgical resection at initial diagnosis, owing to insufficient future liver remnant (FLR) or an inadequate surgical margin. For such patients, portal vein embolization (PVE) is an essential approach to allow liver hypertrophy and prepare for subsequent surgery. However, the conversion resection rate of PVE only is unsatisfactory because of tumor progression while awaiting liver hypertrophy. We report here a successfully treated case of primary massive HCC, where surgical resection was completed after PVE and multimodality therapy, comprising hepatic artery infusion chemotherapy (HAIC), Lenvatinib plus Sintilimab. A pathologic complete response was achieved. This case demonstrates for the first time that combined PVE with multimodality therapy appears to be safe and effective for massive, potentially resectable HCC and can produce deep pathological remission in a primary tumor.

MetaMixUp: Learning Adaptive Interpolation Policy of MixUp With Metalearning.

MixUp is an effective data augmentation method to regularize deep neural networks via random linear interpolations between pairs of samples and their labels. It plays an important role in model regularization, semisupervised learning (SSL), and domain adaption. However, despite its empirical success, its deficiency of randomly mixing samples has poorly been studied. Since deep networks are capable of memorizing the entire data set, the corrupted samples generated by vanilla MixUp with a badly chosen interpolation policy will degrade the performance of networks. To overcome overfitting to corrupted samples, inspired by metalearning (learning to learn), we propose a novel technique of learning to a mixup in this work, namely, MetaMixUp. Unlike the vanilla MixUp that samples interpolation policy from a predefined distribution, this article introduces a metalearning-based online optimization approach to dynamically learn the interpolation policy in a data-adaptive way (learning to learn better). The validation set performance via metalearning captures the noisy degree, which provides optimal directions for interpolation policy learning. Furthermore, we adapt our method for pseudolabel-based SSL along with a refined pseudolabeling strategy. In our experiments, our method achieves better performance than vanilla MixUp and its variants under SL configuration. In particular, extensive experiments show that our MetaMixUp adapted SSL greatly outperforms MixUp and many state-of-the-art methods on CIFAR-10 and SVHN benchmarks under the SSL configuration.

Predicting in vitro single-neuron firing rates upon pharmacological perturbation using Graph Neural Networks.

Modern Graph Neural Networks (GNNs) provide opportunities to study the determinants underlying the complex activity patterns of biological neuronal networks. In this study, we applied GNNs to a large-scale electrophysiological dataset of rodent primary neuronal networks obtained by means of high-density microelectrode arrays (HD-MEAs). HD-MEAs allow for long-term recording of extracellular spiking activity of individual neurons and networks and enable the extraction of physiologically relevant features at the single-neuron and population level. We employed established GNNs to generate a combined representation of single-neuron and connectivity features obtained from HD-MEA data, with the ultimate goal of predicting changes in single-neuron firing rate induced by a pharmacological perturbation. The aim of the main prediction task was to assess whether single-neuron and functional connectivity features, inferred under baseline conditions, were informative for predicting changes in neuronal activity in response to a perturbation with Bicuculline, a GABA A receptor antagonist. Our results suggest that the joint representation of node features and functional connectivity, extracted from a baseline recording, was informative for predicting firing rate changes of individual neurons after the perturbation. Specifically, our implementation of a GNN model with inductive learning capability (GraphSAGE) outperformed other prediction models that relied only on single-neuron features. We tested the generalizability of the results on two additional datasets of HD-MEA recordings-a second dataset with cultures perturbed with Bicuculline and a dataset perturbed with the GABA A receptor antagonist Gabazine. GraphSAGE models showed improved prediction accuracy over other prediction models. Our results demonstrate the added value of taking into account the functional connectivity between neurons and the potential of GNNs to study complex interactions between neurons.

ESCO2 promotes lung adenocarcinoma progression by regulating hnRNPA1 acetylation.

BACKGROUND: Emerging evidence indicates that metabolism reprogramming and abnormal acetylation modification play an important role in lung adenocarcinoma (LUAD) progression, although the mechanism is largely unknown. METHODS: Here, we used three public databases (Oncomine, Gene Expression Omnibus [GEO], The Cancer Genome Atlas [TCGA]) to analyze ESCO2 (establishment of cohesion 1 homolog 2) expression in LUAD. The biological function of ESCO2 was studiedusing cell proliferation, colony formation, cell migration, and invasion assays in vitro, and mouse xenograft models in vivo. ESCO2 interacting proteins were searched using gene set enrichment analysis (GSEA) and mass spectrometry. Pyruvate kinase M1/2 (PKM) mRNA splicing assay was performed using RT-PCR together with restriction digestion. LUAD cell metabolism was studied using glucose uptake assays and lactate production. ESCO2 expression was significantly upregulated in LUAD tissues, and higher ESCO2 expression indicated worse prognosis for patients with LUAD. RESULTS: We found that ESCO2 promoted LUAD cell proliferation and metastasis metabolic reprogramming in vitro and in vivo. Mechanistically, ESCO2 increased hnRNPA1 (heterogeneous nuclear ribonucleoprotein A1) binding to the intronic sequences flanking exon 9 (EI9) of PKM mRNA by inhibiting hnRNPA1 nuclear translocation, eventually inhibiting PKM1 isoform formation and inducing PKM2 isoform formation. CONCLUSIONS: Our findings confirm that ESCO2 is a key factor in promoting LUAD malignant progression and suggest that it is a new target for treating LUAD.

Sophocarpine ameliorates cardiac hypertrophy through activation of autophagic responses.

Mounting evidences indicate that autophagy is an essential homeostatic mechanism to maintain the global cardiac structure function. Sophocarpine (SOP), a major bioactive compound derived from the natural plant Sophora flavescens. However, the role of SOP in cardiac hypertrophy remain to be fully elucidated. In the present study, we tested the hypothesis that SOP protects against Ang II-induced cardiac hypertrophy by mediating the regulation of autophagy. The results demonstrated that SOP attenuated the Ang II-induced cardiac hypertrophy, as assessed by measurements of echocardiography parameters, the ratios of heart weight/body weight and left ventricle weight/body weight, histopathological staining, cross-sectional cardiomyocyte area, and the expression levels of cardiac hypertrophic markers. The anti-hypertrophic effect of SOP was mediated by activating autophagy-related pathway, as revealed by reversal of the increased autophagy marker protein expression. These findings reveal a novel mechanism of SOP attenuating cardiac hypertrophy via activating autophagy-related signaling pathways.

Agmatinase promotes the lung adenocarcinoma tumorigenesis by activating the NO-MAPKs-PI3K/Akt pathway.

Lung adenocarcinoma (LUAD) is one of the leading causes of cancer-related death worldwide. There is an urgent need to uncover the pathogenic mechanism to develop new treatments. Agmatinase (AGMAT) expression and its association with clinicopathological characteristics were analyzed via GEO, Oncomine, and TCGA databases, and IHC staining in human LUAD specimens. An EdU cell proliferation kit, propidiumiodide staining, colony formation, cell migration, and invasion assays, and a xenograft tumor model were used to detect the biological function of AGMAT in LUAD. Furthermore, the expression level of nitric oxide (NO) was detected using a DAF-FMDA fluorescent probe or nitrite assay kit, and further validated with Carboxy-PTIO (a NO scavenger). The roles of three isoforms of nitric oxide synthases (nNOS, eNOS, and iNOS) were validated using L-NAME (eNOS inhibitor), SMT (iNOS inhibitor), and spermidine (nNOS inhibitor). AGMAT expression was up-regulated in LUAD tissues. LUAD patients with high AGMAT levels were associated with poorer prognoses. AGMAT promoted LUAD tumorigenesis in NO released by iNOS both in vitro and in vivo. Importantly, NO signaling up-regulated the expression of cyclin D1 via activating the MAPK and PI3K/Akt-dependent c-myc activity, ultimately promoting the malignant proliferation of tumor cells. On the whole, AGMAT promoted NO release via up-regulating the expression of iNOS. High levels of NO drove LUAD tumorigenesis via activating MAPK and PI3K/Akt cascades. AGMAT might be a potential diagnostic and therapeutic target for LUAD patients.

Biological sequence modeling with convolutional kernel networks.

MOTIVATION: The growing number of annotated biological sequences available makes it possible to learn genotype-phenotype relationships from data with increasingly high accuracy. When large quantities of labeled samples are available for training a model, convolutional neural networks can be used to predict the phenotype of unannotated sequences with good accuracy. Unfortunately, their performance with medium- or small-scale datasets is mitigated, which requires inventing new data-efficient approaches. RESULTS: We introduce a hybrid approach between convolutional neural networks and kernel methods to model biological sequences. Our method enjoys the ability of convolutional neural networks to learn data representations that are adapted to a specific task, while the kernel point of view yields algorithms that perform significantly better when the amount of training data is small. We illustrate these advantages for transcription factor binding prediction and protein homology detection, and we demonstrate that our model is also simple to interpret, which is crucial for discovering predictive motifs in sequences. AVAILABILITY AND IMPLEMENTATION: Source code is freely available at https://gitlab.inria.fr/dchen/CKN-seq. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Overexpression of TPX2 is associated with progression and prognosis of prostate cancer.

Targeting protein for Xenopus kinesin-like protein 2 (TPX2) activates Aurora kinase A during mitosis and targets its activity to the mitotic spindle, serving an important role in mitosis. It has been associated with different types of cancer and is considered to promote tumor growth. The aim of the present study was to explore the role of TPX2 in diagnosing prostate cancer (PCa). It was identified that TPX2 expression in PCa tissues was increased compared with benign prostate tissues. Microarray analysis demonstrated that TPX2 was positively associated with the Gleason score, tumor-node-metastasis (TNM) stage, clinicopathological stage, metastasis, overall survival and biochemical relapse-free survival. In vitro studies revealed that the high expression of TPX2 in PCa cells improved proliferative, invasive and migratory abilities, and repressed apoptosis of the PCa cells, without affecting tolerance to docetaxel. The results suggested that TPX2 serves as a tumorigenesis-promoting gene in PCa, and a potential therapeutic target for patients with PCa.