Sparse discriminant PCA based on contrastive learning and class-specificity distribution.

Much mathematical effort has been devoted to developing Principal Component Analysis (PCA), which is the most popular feature extraction method. To suppress the negative effect of noise on PCA performance, there have been extensive studies and applications of a large number of robust PCAs achieving outstanding results. However, existing methods suffer from at least two shortcomings: (1) They expressed PCA as a reconstruction model measured by Euclidean distance, which only considers the relationship between the data and its reconstruction and ignores the differences between different data points; (2) They did not consider the class-specificity distribution information contained in the data itself, thus lacking discriminative properties. To overcome the above problems, we propose a Sparse Discriminant Principal Components Analysis (SDPCA) model based on contrastive learning and class-specificity distribution. Specifically, we use contrastive learning to measure the relationship between samples and their reconstructions, which fully takes the discriminative information between data into account in PCA. In order to make the extracted low-dimensional features profoundly reflect the class-specificity distribution of the data, we minimize the squared ℓ1,2-norm of the low-dimensional embedding. In addition, to reduce the effects of redundant features and noise and to improve the interpretability of PCA at the same time, we impose sparsity constraints on the projection matrix using the squared ℓ1,2-norm. Our experimental results on different types of benchmark databases demonstrate that our model has state-of-the-art performance.

Different Levels of Context-Specificity of Teacher Self-Efficacy and Their Relations With Teaching Quality.

On the basis of Bandura's social cognitive theory, researchers often assume that a teachers' self-efficacy (TSE) will have a positive effect on teaching quality. However, the available empirical evidence is mixed. Building on previous research into TSE, we examined whether assessing class-/task-specific TSE gives a more accurate indication of the associations between TSE assessments and student-rated teaching quality. The analyses were based on the English sample of the Teaching and Learning International Survey (TALIS) Video Study. Mathematics teachers (N = 86) rated their self-efficacy beliefs using generalized task-specific TSE items and class-/task-specific TSE items. Their students (N = 1,930) rated the quality of teaching in their math class. Multilevel regression analyses revealed stronger associations between student-rated teaching quality and class-/task-specific TSE than generalized task-specific TSE. We discuss possible reasons for these results and outline the potential benefits of using class-specific assessments for future TSE research.

Two binding epitopes modulating specificity of immunoassay for ß-agonist detection: Quantitative structure-activity relationship.

A growing number of ß-agonists are illegally using for reducing animal fat deposition in animals, but the development of analytical methods always lags behind the emergence of new illegal compounds. Therefore, class specificity antibody-based immunoassays that can detect a great many ß-agonists are important for timely supervision. In this study, a competitive inhibition enzyme-linked immunosorbent assay (ciELISA) based on a clenbuterol monoclonal antibody was developed to recognize 23 ß-agonists and analogues. Holographic and three-dimensional quantitative structure-activity relationship (HQSAR and 3D QSAR) revealed that there are two critical binding epitopes on ß-agonist hapten affecting antibody specificity, and these epitopes have been further validated using a ractopamine antibody with narrow specificity. Tert-butyl at C-2' epitope is needed to generate class specific antibodies, and different characteristics of substituents at benzene ring epitope would adjust antibody specificity. This investigation could provide reference for future design of ß-agonist haptens.

Preparation and Characterization of Monoclonal Antibodies with High Affinity and Broad Class Specificity against Zearalenone and Its Major Metabolites.

This study aimed to detect and monitor total Zearalenone (ZEN) and its five homologs (ZENs) in cereals and feed. The monoclonal antibodies (mAbs) with a high affinity and broad class specificity against ZENs were prepared, and the conditions of a heterologous indirect competitive ELISA (icELISA) were preliminarily optimized based on the ZEN mAbs. The immunogen ZEN-BSA was synthesized using the oxime active ester method (OAE) and identified using infrared (IR) and ultraviolet (UV). The coating antigen ZEN-OVA was obtained via the 1,4-butanediol diglycidyl ether method (BDE). Balb/c mice were immunized using a high ZEN-BSA dose with long intervals and at multiple sites. A heterologous indirect non-competitive ELISA (inELISA) and an icELISA were used to screen the suitable cell fusion mice and positive hybridoma cell lines. The ZEN mAbs were prepared by inducing ascites in vivo. The standard curve was established, and the sensitivity and specificity of the ZEN mAbs were determined under the optimized icELISA conditions. ZEN-BSA was successfully synthesized at a conjugation ratio of 17.2:1 (ZEN: BSA). Three hybridoma cell lines, 2D7, 3C2, and 4A10, were filtered, and their mAbs corresponded to an IgG1 isotype with a κ light chain. The mAbs titers were between (2.56 to 5.12) × 102 in supernatants and (1.28 to 5.12) × 105 in the ascites. Besides, the 50% inhibitive concentration (IC50) values were from 18.65 to 31.92 µg/L in the supernatants and 18.12 to 31.46 µg/L in the ascites. The affinity constant (Ka) of all of the mAbs was between 4.15 × 109 and 6.54 × 109 L/mol. The IC50 values of mAb 2D7 for ZEN, α-ZEL, ß-ZEL, α-ZAL, ß-ZAL and ZAN were 17.23, 16.71, 18.27, 16.39, 20.36 and 15.01 µg/L, and their cross-reactivities (CRs, %) were 100%, 103.11%, 94.31%, 105.13%, 84.63%, and 114.79%, respectively, under the optimized icELISA conditions. The limit of detection (LOD) for ZEN was 0.64 µg/L, and its linear working range was between 1.03 and 288.55 µg/L. The mAbs preparation and the optimization of icELISA conditions promote the potential development of a rapid test ELISA kit, providing an alternative method for detecting ZEN and its homologs in cereals and feed.

Domain swaps of Arabidopsis secondary wall cellulose synthases to elucidate their class specificity.

Cellulose microfibrils are synthesized by membrane-embedded cellulose synthesis complexes (CSCs), currently modeled as hexamers of cellulose synthase (CESA) trimers. The three paralogous CESAs involved in secondary cell wall (SCW) cellulose biosynthesis in Arabidopsis (CESA4, CESA7, CESA8) are similar, but nonredundant, with all three isoforms required for assembly and function of the CSC. The molecular basis of protein-protein recognition among the isoforms is not well understood. To investigate the locations of the interfaces that are responsible for isoform recognition, we swapped three domains between the Arabidopsis CESAs required for SCW synthesis (CESA4, CESA7, and CESA8): N-terminus, central domain containing the catalytic core, and C-terminus. Chimeric genes with all pairwise permutations of the domains were tested for in vivo functionality within knockout mutant backgrounds of cesa4, cesa7, and cesa8. Immunoblotting with isoform-specific antibodies confirmed the anticipated protein expression in transgenic plants. The percent recovery of stem height and crystalline cellulose content was assayed, as compared to wild type, the mutant background lines, and other controls. Retention of the native central domain was sufficient for CESA8 chimeras to function, with neither its N-terminal nor C-terminal domains required. The C-terminal domain is required for class-specific function of CESA4 and CESA7, and CESA7 also requires its own N-terminus. Across all isoforms, the results indicate that the central domain, as well as the N- and C-terminal regions, contributes to class-specific function variously in Arabidopsis CESA4, CESA7, and CESA8.

Characterisation of a type 1 Avian Paramyxovirus belonging to a divergent group.

Newcastle disease, induced by a type 1 Avian Paramyxovirus (APMV-1), is one of the most serious poultry diseases. APMV-1 are divided into two classes based on genetic analysis: class II strains have been recovered from wild or domestic birds and include virulent and avirulent isolates whereas class I strains have been mainly isolated from wild birds and are avirulent. Within class I, a new proposed genotype has recently been reported. The only full genome strain of this group is presently characterised from the point of view of codon usage with reference to class I and class II specificities. Class-specific residues were identified on HN and F proteins that are the two major proteins involved in cell attachment and pathogenicity. Comparison of protein patterns and codon usage for this newly identified APMV-1 strain indicates it is similar to class I viruses but contains a few characteristics close to the viruses of class II. Transmission of viruses from this recently identified divergent group from wild birds to domestic birds could have a major impact on the domestic poultry industry.