COVID-Net L2C-ULTRA: An Explainable Linear-Convex Ultrasound Augmentation Learning Framework to Improve COVID-19 Assessment and Monitoring.

While no longer a public health emergency of international concern, COVID-19 remains an established and ongoing global health threat. As the global population continues to face significant negative impacts of the pandemic, there has been an increased usage of point-of-care ultrasound (POCUS) imaging as a low-cost, portable, and effective modality of choice in the COVID-19 clinical workflow. A major barrier to the widespread adoption of POCUS in the COVID-19 clinical workflow is the scarcity of expert clinicians who can interpret POCUS examinations, leading to considerable interest in artificial intelligence-driven clinical decision support systems to tackle this challenge. A major challenge to building deep neural networks for COVID-19 screening using POCUS is the heterogeneity in the types of probes used to capture ultrasound images (e.g., convex vs. linear probes), which can lead to very different visual appearances. In this study, we propose an analytic framework for COVID-19 assessment able to consume ultrasound images captured by linear and convex probes. We analyze the impact of leveraging extended linear-convex ultrasound augmentation learning on producing enhanced deep neural networks for COVID-19 assessment, where we conduct data augmentation on convex probe data alongside linear probe data that have been transformed to better resemble convex probe data. The proposed explainable framework, called COVID-Net L2C-ULTRA, employs an efficient deep columnar anti-aliased convolutional neural network designed via a machine-driven design exploration strategy. Our experimental results confirm that the proposed extended linear-convex ultrasound augmentation learning significantly increases performance, with a gain of 3.9% in test accuracy and 3.2% in AUC, 10.9% in recall, and 4.4% in precision. The proposed method also demonstrates a much more effective utilization of linear probe images through a 5.1% performance improvement in recall when such images are added to the training dataset, while all other methods show a decrease in recall when trained on the combined linear-convex dataset. We further verify the validity of the model by assessing what the network considers to be the critical regions of an image with our contribution clinician.

DNA methylation differences in monozygotic twins with Van der Woude syndrome.

Introduction: Van der Woude Syndrome (VWS) is an autosomal dominant disorder responsible for 2% of all syndromic orofacial clefts (OFCs) with IRF6 being the primary causal gene (70%). Cases may present with lip pits and either cleft lip, cleft lip with cleft palate, or cleft palate, with marked phenotypic discordance even among individuals carrying the same mutation. This suggests that genetic or epigenetic modifiers may play additional roles in the syndrome's etiology and variability in expression. We report the first DNA methylation profiling of 2 pairs of monozygotic twins with VWS. Our goal is to explore epigenetic contributions to VWS etiology and variable phenotypic expressivity by comparing DNAm profiles in both twin pairs. While the mutations that cause VWS in these twins are known, the additional mechanism behind their phenotypic risk and variability in expression remains unclear. Methods: We generated whole genome DNAm data for both twin pairs. Differentially methylated positions (DMPs) were selected based on: (1) a coefficient of variation in DNAm levels in unaffected individuals < 20%, and (2) intra-twin pair absolute difference in DNAm levels >5% (delta beta > | 0.05|). We then divided the DMPs in two subgroups for each twin pair for further analysis: (1) higher methylation levels in twin A (Twin A > Twin B); and (2) higher methylation levels in twin B (Twin B >Twin A). Results and Discussion: Gene ontology analysis revealed a list of enriched genes that showed significant differential DNAm, including clef-associated genes. Among the cleft-associated genes, TP63 was the most significant hit (p=7.82E-12). Both twin pairs presented differential DNAm levels in CpG sites in/near TP63 (Twin 1A > Twin 1B and Twin 2A < Twin 2B). The genes TP63 and IRF6 function in a biological regulatory loop to coordinate epithelial proliferation and differentiation in a process that is critical for palatal fusion. The effects of the causal mutations in IRF6 can be further impacted by epigenetic dysregulation of IRF6 itself, or genes in its pathway. Our data shows evidence that changes in DNAm is a plausible mechanism that can lead to markedly distinct phenotypes, even among individuals carrying the same mutation.

TRAF3IP2-IL-17 Axis Strengthens the Gingival Defense against Pathogens.

Recent genome-wide association studies have suggested novel risk loci associated with periodontitis, which is initiated by dysbiosis in subgingival plaque and leads to destruction of teeth-supporting structures. One such genetic locus was the tumor necrosis factor receptor-associated factor 3 interacting protein 2 (TRAF3IP2), a gene encoding the gate-keeping interleukin (IL)-17 receptor adaptor. In this study, we first determined that carriers of the lead exonic variant rs13190932 within the TRAF3IP2 locus combined with a high plaque microbial burden was associated with more severe periodontitis than noncarriers. We then demonstrated that TRAF3IP2 is essential in the IL-17-mediated CCL2 and IL-8 chemokine production in primary gingival epithelial cells. Further analysis suggested that rs13190932 may serve a surrogate variant for a genuine loss-of-function variant rs33980500 within the same gene. Traf3ip2 null mice (Traf3ip2-/-) were more susceptible than wild-type (WT) mice to the Porphyromonas gingivalis-induced periodontal alveolar bone loss. Such bone loss was associated with a delayed P. gingivalis clearance and an attenuated neutrophil recruitment in the gingiva of Traf3ip2-/- mice. Transcriptomic data showed decreased expression of antimicrobial genes, including Lcn2, S100a8, and Defb1, in the Traf3ip2-/- mouse gingiva in comparison to WT mice prior to or upon P. gingivalis oral challenge. Further 16S ribosomal RNA sequencing analysis identified a distinct microbial community in the Traf3ip2-/- mouse oral plaque, which was featured by a reduced microbial diversity and an overabundance of Streptococcus genus bacteria. More P. gingivalis was observed in the Traf3ip2-/- mouse gingiva than WT control animals in a ligature-promoted P. gingivalis invasion model. In agreement, neutrophil depletion resulted in more local gingival tissue invasion by P. gingivalis. Thus, we identified a homeostatic IL-17-TRAF3IP2-neutrophil axis underpinning host defense against a keystone periodontal pathogen.

Genome-wide Gene-by-Sex Interaction Studies Identify Novel Nonsyndromic Orofacial Clefts Risk Locus.

Risk loci identified through genome-wide association studies have explained about 25% of the phenotypic variations in nonsyndromic orofacial clefts (nsOFCs) on the liability scale. Despite the notable sex differences in the incidences of the different cleft types, investigation of loci for sex-specific effects has been understudied. To explore the sex-specific effects in genetic etiology of nsOFCs, we conducted a genome-wide gene × sex (GxSex) interaction study in a sub-Saharan African orofacial cleft cohort. The sample included 1,019 nonsyndromic orofacial cleft cases (814 cleft lip with or without cleft palate and 205 cleft palate only) and 2,159 controls recruited from 3 sites (Ethiopia, Ghana, and Nigeria). An additive logistic model was used to examine the joint effects of the genotype and GxSex interaction. Furthermore, we examined loci with suggestive significance (P < 1E-5) in the additive model for the effect of the GxSex interaction only. We identified a novel risk locus on chromosome 8p22 with genome-wide significant joint and GxSex interaction effects (rs2720555, p2df = 1.16E-08, pGxSex = 1.49E-09, odds ratio [OR] = 0.44, 95% CI = 0.34 to 0.57). For males, the risk of cleft lip with or without cleft palate at this locus decreases with additional copies of the minor allele (p < 0.0001, OR = 0.60, 95% CI = 0.48 to 0.74), but the effect is reversed for females (p = 0.0004, OR = 1.36, 95% CI = 1.15 to 1.60). We replicated the female-specific effect of this locus in an independent cohort (p = 0.037, OR = 1.30, 95% CI = 1.02 to 1.65), but no significant effect was found for the males (p = 0.29, OR = 0.86, 95% CI = 0.65 to 1.14). This locus is in topologically associating domain with craniofacially expressed and enriched genes during embryonic development. Rare coding mutations of some of these genes were identified in nsOFC cohorts through whole exome sequencing analysis. Our study is additional proof that genome-wide GxSex interaction analysis provides an opportunity for novel findings of loci and genes that contribute to the risk of nsOFCs.

The lncRNA HOXA11-AS promotes glioma cell growth and metastasis by targeting miR-130a-5p/HMGB2.

OBJECTIVE: Long noncoding RNAs (lncRNAs) serve as important regulators of diverse types of cancer, including glioma. Nevertheless, their precise roles in cancers remain sufficiently unexplored. PATIENTS AND METHODS: Quantitative Reverse Transcription-Polymerase Chain Reaction (qRT-PCR) was used to determine the levels of HOMEOBOX A11 antisense RNA (HOXA11-AS) and miR-130a-5p in glioma tissues and cell lines. Short hairpin RNAs (shRNAs) targeting HOXA11-AS or pcDNA3.1 were transfected into cells via a vector encoding HOXA11-AS to decrease or increase the level of HOXA11-AS. Cell Counting Kit-8 (CCK-8), colony formation, wound healing, flow cytometry and transwell assays were applied to assess the role of HOXA11-AS in glioblastoma cell growth, apoptosis and aggressiveness. The expression of N-cadherin and E-cadherin was determined using immunofluorescence staining. The expression of high-mobility group protein B2 (HMGB2) was determined using Western blot analysis in vitro and immunohistochemistry (IHC) staining in vivo. The direct target of HOXA11-AS and miR-130a-5p was confirmed using the Luciferase reporter assay. Glioblastoma cells were subcutaneously implanted into nude mice to determine the role of HOXA11-AS in tumor growth in vivo. RESULTS: In the current study, we demonstrated that the lncRNA HOXA11-AS was overexpressed in glioma. The overexpression of HOXA11-AS was correlated with advanced stages of glioma and poor prognosis. Downregulating HOXA11-AS expression significantly suppressed the proliferation, migration and invasion of glioma cells and increased their apoptosis. The growth of glioma cells in vitro was also suppressed by the downregulation of HOXA11-AS. Finally, we revealed that HOXA11-AS exerted its oncogenic effects by binding to miR-130a-5p, thereby neutralizing the suppressive effect of miR-130a-5p on HMGB2. CONCLUSIONS: Our results demonstrate that HOXA11-AS regulates the growth and metastasis of glioma by targeting the miR-130a-5p-HMGB2 signaling axis.

Pharmacokinetics of LJP 993, a tetrameric conjugate of domain 1 of beta2-glycoprotein I for antiphospholipid syndrome.

beta2-glycoprotein I is the best-characterized antigenic target for antiphospholipid autoantibodies. We synthesized a tetrameric conjugate of the domain 1 of beta2-glycoprotein I (LJP 993) aimed at developing the conjugate as a Toleragen to suppress antiphospholipid syndrome. The present studies focused on determining the stability, tissue distribution, plasma concentration-time profile and excretion of the LJP 993 in mice. The stability of LJP 993 in mouse plasma was quantitatively evaluated using strong cation-exchange high performance liquid chromatography. ( 125)I-labeled LJP 993 was intravenously injected to mice, and levels of (125)I-labeled LJP 993 in plasma, tissues, urine and feces were determined at known intervals. Incubation of LJP 993 with mouse serum at 37 degrees C for 8 h resulted in a decrease by 34% of LJP 993 concentration. No degradation fragment was observed during the incubation. After a single intravenous administration of (125)I-LJP 993 (0.5 and 5 mg/kg) to mice, both C(max) and area-under-curve values increased in a dose-proportional manner, and blood radioactivity disappeared in a bi-exponential manner with the distribution half-lives equal to 1.7 min, and the elimination half-lives 188 and 281 min, respectively. The (125)I-LJP 993 was moderately distributed into organs and tissues with the exception that brain level of ( 125)I-LJP 993 was negligible. The major sites of (125)I-LJP 993 uptake were the kidney (at 30 min post dosing), and kidney, lung, liver, heart, spleen, skin, muscle and fat tissues (at 4 h post dosing). Cumulative urinary and fecal radioactivity for 0-48 h post dosing accounted for 44.7% and 4.2% of the administered dose, respectively, with the fast rate of urinal excretion occurring within the first 8 h. In summary, LJP 993 was fairly stable in mouse plasma. After administration to mice, (125)I-LJP 993 was taken up mainly by kidney and then distributed extensively to tissues except brain. Both C(max) and area-under-curve values increased in a dose-proportional manner. It was predominantly excreted in the urine with an elimination half-life longer than 3 h. Kidney is a major route to excrete the tetrameric conjugate.

[The nucleic acid analysis by DNA chip technique based on nuclease S1 protection].

OBJECTIVE: Establishing a method for quantitative analysis of nucleic acid by DNA chips. METHODS: The modified oligonucleotides with ribose at 3'-end was chemically synthesized. The 5'-end was labeled by radioisotope 32P with kinase catalyzed reactions. Such oligonucleotides were converted into di-aldehyde at 3'-end by oxidization with NaIO4, and then were spotted on glass slide with the amino group modified surface. After reduced with NaBH4, the oligonucleotides were attached strongly. The DNA chips prepared with this method were hybridized with nucleic acids existed in the solution and then digested with nuclease S1. RESULTS: When they were paired with the nucleic acids in the solution perfectly, the oligonucleotides on the chip were not cleaved by nuclease S1. Otherwise, the oligonucleotides on chip were cleaved. The protection efficiencies appeared proportional to the perfect paired nucleic acids in the solution when the content of target nucleic acids were less than the spots on the slides. CONCLUSIONS: The method was developed for both qualitative and quantitative analysis of nucleic acid. As it was not required to label the samples with radioisotope or fluorescence, it might be a practical choice for clinical tests.

Butyltins in sediments from Santa Monica and San Pedro basins, California.

Butyltins have been measured for the first time in sediments from the deeper waters of the Santa Monica and San Pedro (SM/SP) basins of Southern California borderland. Core samples were collected from water depths ranging from 458 m in the shelf to 906 m in the central basin. Surficial and a few subsurface sections from selected cores were analyzed for butyltins. The dominant components are generally dibutyl and monobutyltins (DBT and MBT, respectively). Microbial degradation of Tributyltin (TBT, the most toxic of the butyltin species) to DBT and MBT during the long range transport to deeper basins, as well as a lack of continuing inputs of TBT in the present times could justify the butyltins signatures found in the region. The levels of butyltins are in the lower range (below detection level, <1, to 14 ng of Sn g(-1)) relative to nearshore sediments. However, the results demonstrate that the butyltins are, indeed, found even in the sediments of the central basins of the Southern California borderland suggesting that deeper global oceanic regimes need to be further investigated for these compounds.