Novel Method for Three-Dimensional Facial Expression Recognition Using Self-Normalizing Neural Networks and Mobile Devices

Introduction To date, most ways to perform facial expression recognition rely on two-dimensional images, advanced approaches with three-dimensional data exist. These however demand stationary apparatuses and thus lack portability and possibilities to scale deployment. As human emotions, intent and even diseases may condense in distinct facial expressions or changes therein, the need for a portable yet capable solution is signified. Due to the superior informative value of three-dimensional data on facial morphology and because certain syndromes find expression in specific facial dysmorphisms, a solution should allow portable acquisition of true three-dimensional facial scans in real time. In this study we present a novel solution for the three-dimensional acquisition of facial geometry data and the recognition of facial expressions from it. The new technology presented here only requires the use of a smartphone or tablet with an integrated TrueDepth camera and enables real-time acquisition of the geometry and its categorization into distinct facial expressions. Material and Methods Our approach consisted of two parts: First, training data was acquired by asking a collective of 226 medical students to adopt defined facial expressions while their current facial morphology was captured by our specially developed app running on iPads, placed in front of the students. In total, the list of the facial expressions to be shown by the participants consisted of “disappointed”, “stressed”, “happy”, “sad” and “surprised”. Second, the data were used to train a self-normalizing neural network. A set of all factors describing the current facial expression at a time is referred to as “snapshot”. Results In total, over half a million snapshots were recorded in the study. Ultimately, the network achieved an overall accuracy of 80.54% after 400 epochs of training. In test, an overall accuracy of 81.15% was determined. Recall values differed by the category of a snapshot and ranged from 74.79% for “stressed” to 87.61% for “happy”. Precision showed similar results, whereas “sad” achieved the lowest value at 77.48% and “surprised” the highest at 86.87%. Conclusions With the present work it can be demonstrated that respectable results can be achieved even when using data sets with some challenges. Through various measures, already incorporated into an optimized version of our app, it is to be expected that the training results can be significantly improved and made more precise in the future. Currently a follow-up study with the new version of our app that encompasses the suggested alterations and adaptions, is being conducted. We aim to build a large and open database of facial scans not only for facial expression recognition but to perform disease recognition and to monitor diseases’ treatment progresses.     

Receptor-Binding Domain (RBD) Antibodies Contribute More to SARS-CoV-2 Neutralization When Target Cells Express High Levels of ACE2

Neutralization assays are experimental surrogates for the effectiveness of infection- or vaccine-elicited polyclonal antibodies and therapeutic monoclonal antibodies targeting SARS-CoV-2. However, the measured neutralization can depend on the details of the experimental assay. Here, we systematically assess how ACE2 expression in target cells affects neutralization by antibodies to different spike epitopes in lentivirus pseudovirus neutralization assays. For high ACE2-expressing target cells, receptor-binding domain (RBD) antibodies account for nearly all neutralizing activity in polyclonal human sera. However, for lower ACE2-expressing target cells, antibodies targeting regions outside the RBD make a larger (although still modest) contribution to serum neutralization. These serum-level results are mirrored for monoclonal antibodies: N-terminal domain (NTD) antibodies and RBD antibodies that do not compete for ACE2 binding incompletely neutralize on high ACE2-expressing target cells, but completely neutralize on cells with lower ACE2 expression. Our results show that the ACE2 expression level in the target cells is an important experimental variable, and that high ACE2 expression emphasizes the role of a subset of RBD-directed antibodies.     

Gene Expression Analysis of Adapted Insect Cells during Influenza VLP Production Using RNA-Sequencing

Adaptive laboratory evolution has been used to improve production of influenza hemagglutinin (HA)-displaying virus-like particles (VLPs) in insect cells. However, little is known about the underlying biological mechanisms promoting higher HA-VLP expression in such adapted cell lines. In this article, we present a study of gene expression patterns associated with high-producer insect High Five cells adapted to neutral pH, in comparison to non-adapted cells, during expression of influenza HA-VLPs. RNA-seq shows a decrease in the amount of reads mapping to host cell genomes along infection, and an increase in those mapping to baculovirus and transgenes. A total of 1742 host cell genes were found differentially expressed between adapted and non-adapted cells throughout infection, 474 of those being either up- or down-regulated at both time points evaluated (12 and 24 h post-infection). Interestingly, while host cell genes were found up- and down-regulated in an approximately 1:1 ratio, all differentially expressed baculovirus genes were found to be down-regulated in infected adapted cells. Pathway analysis of differentially expressed genes revealed enrichment of ribosome biosynthesis and carbohydrate, amino acid, and lipid metabolism. In addition, oxidative phosphorylation and protein folding, sorting and degradation pathways were also found to be overrepresented. These findings contribute to our knowledge of biological mechanisms of insect cells during baculovirus-mediated transient expression and will assist the identification of potential engineering targets to increase recombinant protein production in the future.     

HE4 Tissue Expression as A Putative Prognostic Marker in Low-Risk/Low-Grade Endometrioid Endometrial Cancer: A Review

Low-grade stage I endometrioid endometrial carcinomas should have an excellent prognosis, but a small subset of these cancers can relapse. The search for putative immunohistochemical prognostic markers for relapse in low-risk/low-grade endometrioid endometrial cancers remains open. Among the candidate molecules that may implicate the roles of immunohistochemical risk markers, we focused our attention on human epididymis protein 4 (HE4) after a review of the literature. Few authors have devoted themselves to this topic, and none have found a correlation between the tissue expression of HE4 and the molecular classification of endometrial cancer. Five different variants of HE4 mRNA and multiple protein isoforms of HE4 were identified many years ago, but current HE4 assays only measure the total HE4 expression and do not distinguish the different proteins encoded by different mRNA variants. It is important to have an approach to distinguish specific variants in the future.     

Gene Expression Signatures of Contact Lens-Induced Myopia in Guinea Pig Retinal Pigment Epithelium

PurposeTo identify key retinal pigment epithelium (RPE) genes linked to the induction of myopia in guinea pigs.MethodsTo induce myopia, two-week-old pigmented guinea pigs (New Zealand strain, n = 5) wore −10 diopter (D) rigid gas-permeable contact lenses (CLs), for one day; fellow eyes were left without CLs and served as controls. Spherical equivalent refractive errors (SE) and axial length (AL) were measured at baseline and one day after initiation of CL wear. RNA sequencing was applied to RPE collected from both treated and fellow (control) eyes after one day of CL-wear to identify related gene expression changes. Additional RPE-RNA samples from treated and fellow eyes were subjected to quantitative real-time PCR (qRT-PCR) analysis for validation purposes.ResultsThe CLs induced myopia. The change from baseline values in SE was significantly different (P = 0.016), whereas there was no significant difference in the change in AL (P = 0.10). RNA sequencing revealed significant interocular differences in the expression in RPE of 13 genes: eight genes were significantly upregulated in treated eyes relative to their fellows, and five genes, including bone morphogenetic protein 2 (Bmp2), were significantly downregulated. The latter result was also confirmed by qRT-PCR. Additional analysis of differentially expressed genes revealed significant enrichment for bone morphogenetic protein (BMP) and TGF-β signaling pathways.ConclusionsThe results of this RPE gene expression study provide further supporting evidence for an important role of BMP2 in eye growth regulation, here from a guinea pig myopia model.     

Improved Expression of SARS-CoV-2 Spike RBD Using the Insect Cell-Baculovirus System

Insect cell-baculovirus expression vector system is one of the most established platforms to produce biological products, and it plays a fundamental role in the context of COVID-19 emergency, providing recombinant proteins for treatment, diagnosis, and prevention. SARS-CoV-2 infection is mediated by the interaction of the spike glycoprotein trimer via its receptor-binding domain (RBD) with the host’s cellular receptor. As RBD is required for many applications, in the context of pandemic it is important to meet the challenge of producing a high amount of recombinant RBD (rRBD). For this reason, in the present study, we developed a process based on Sf9 insect cells to improve rRBD yield. rRBD was recovered from the supernatant of infected cells and easily purified by metal ion affinity chromatography, with a yield of 82% and purity higher than 95%. Expressed under a novel chimeric promoter (polh-pSeL), the yield of rRBD after purification was 21.1 ± 3.7 mg/L, which is the highest performance described in Sf9 cell lines. Finally, rRBD was successfully used in an assay to detect specific antibodies in COVID-19 serum samples. The efficient strategy herein described has the potential to produce high-quality rRBD in Sf9 cell line for diagnostic purpose.