A new, to my knowledge, procedure for retrieving the wave aberration from the point-spread function is presented. It uses the Levenberg-Marquardt optimization algorithm in a mutiresolution pyramidal scheme. The method, tested with simulated large aberrations without initial estimates, accelerates convergence and avoids stagnation in local minima. 相似文献
To study the influence of genomic context on transgene expression, we have determined the T-DNA structure, flanking DNA sequences, and chromosomal location of four independent transgene loci in tobacco. Two of these loci were stably expressed in the homozygous condition over many generations, whereas the other two loci became unstable after several generations of homozygosity. The stably expressed loci comprised relatively simple T-DNA arrangements that were flanked on at least one side by plant DNA containing AT-rich regions that bind to nuclear matrices in vitro. Of the unstably expressed loci, one consisted of multiple incomplete T-DNA copies, and the second contained a single intact T-DNA; in both cases, however, binary vector sequences were directly contiguous to a right T-DNA border. Fluorescence in situ hybridization demonstrated that the two stably expressed inserts were present in the vicinity of telomeres. The two unstably expressed inserts occupied intercalary and paracentromeric locations, respectively. Results on the stability of transgene expression in F1 progeny obtained by intercrossing the four lines and the sensitivity of the four transgene loci to inactivation in the presence of an unlinked "trans-silencing" locus are also presented. The findings are discussed in the context of repetitive DNA sequences and the allotetraploid nature of the tobacco genome. 相似文献
Smart devices, such as smartphones, wearables, robots, and others, can collect vast amounts of data from their environment. This data is suitable for training machine learning models, which can significantly improve their behavior, and therefore, the user experience. Federated learning is a young and popular framework that allows multiple distributed devices to train deep learning models collaboratively while preserving data privacy. Nevertheless, this approach may not be optimal for scenarios where data distribution is non-identical among the participants or changes over time, causing what is known as concept drift. Little research has yet been done in this field, but this kind of situation is quite frequent in real life and poses new challenges to both continual and federated learning. Therefore, in this work, we present a new method, called Concept-Drift-Aware Federated Averaging (CDA-FedAvg). Our proposal is an extension of the most popular federated algorithm, Federated Averaging (FedAvg), enhancing it for continual adaptation under concept drift. We empirically demonstrate the weaknesses of regular FedAvg and prove that CDA-FedAvg outperforms it in this type of scenario.
An overview is given on pesticide-residue determination in environmental and food samples by liquid chromatography/mass spectrometry/mass spectrometry (LC/MS/MS). Pesticides comprise a large number of substances that belong to many completely different chemical groups, the only common characteristic is that they are effective against pests. They still constitute a challenge in MS because there is no collective pathway for fragmentation. A brief introduction to the theory of tandem MS permits a discussion of which parameters influence the ionization efficiency when the ions are subjected to different actions. Emphasis is placed on the different tandem MS instruments: triple and ion-trap quadrupoles, and hybrid quadrupole time-of-flight (Q-TOF), including advantages and drawbacks, typical detection limits, and ion signals at low concentrations. The instrumental setup, as well as LC and mass spectrometric experimental conditions, must be carefully selected to increase the performance of the analytical system. The capacity of each instrument to provide useful data for the identification of pesticides, and the possibility to obtain structural information for the identification of target and non-target compounds, are discussed. Finally, sample preparation techniques and examples of applications are debated to reveal the potential of the current state-of-the-art technology, and to further promote the usefulness of tandem MS. 相似文献
Abstract: The paper describes causal network models, giving a detailed discussion of their basic assumptions and their associated algorithms for marginalization and updating. Conditional probability tables are shown to be the basis for the aggregation formulas of if-then rules allowing for deriving the uncertainty of conclusions from that of premises. Examples of a pressure tank and a power system are used for illustrating the concepts and all steps of explained algorithms. 相似文献
Identification of degradation products of the organophophorous pesticide fenthion formed in two orange varieties, Valencia Navel and Navel Late, under field conditions has been assessed using liquid chromatography quadrupole time-of-flight mass spectrometry and ion trap mass spectrometry. The structural elucidation of the metabolites was accomplished by the accurate mass measurements provided by the quadrupole time-of-flight mass spectrometer in MS and MS/MS modes. This instrument achieved elemental composition diagnosis for the precursor and product ions with absolute mass error of <5 ppm, which unambiguously establishes the identity of the metabolites even at low concentration. The presence of these compounds was also confirmed by electrospray ionization-ion trap mass spectrometry, performing successive fragmentation steps (MS(n)). Once identified, each molecule was confirmed by comparison with its analytical standard, also used to explore the quantitative capabilities of both mass analyzers. The extraction method was evaluated because it predetermines the metabolites that can be found (e.g., according to their polarity). Recoveries ranged from 70% for fenoxon sulfoxide (the most polar) to 101% for fenthion (the most apolar), which also indicates the method's facility to extract other more polar metabolites if present. Satisfactory linear range (r > 0.99) of more than 2 orders of magnitude was obtained with both analyzers for standards prepared in methanol and in untreated orange extracts. However, the matrix-matched standards showed suppression of the mass signal due to the matrix effect, especially for fenoxon sulfoxide and sulfone. The limits of quantification ranged from 0.005 to 0.015 mg/kg. The QqTOF-MS provided better quantification limits for fenthion and its sulfoxide and sulfone than the IT-MS. The resulting fenthion degration curves in oranges indicated that it was mainly degraded by sunlight photolysis to its sulfoxide and sulfone. However, hydrolysis was also observed by the appearance of fenoxon, fenoxon sulfoxide, and fenoxon sulfone, but always in low concentrations, which can be related to the rain events. 相似文献
Whether there is one or multiple alphabetaT cell antigen receptor (TCR) recognition modules in a given TCR/CD3 complex is a long-standing controversy in immunology. We show that T cells from transgenic mice that coexpress comparable amounts of two distinct TCRbeta chains incorporate at least two alphabetaTCRs in a single TCR/CD3 complex. Evidence for bispecific alphabetaTCRs was obtained by immunoprecipitation and immunoblotting and confirmed on the surface of living cells both by fluorescence resonance energy transfer and comodulation assays by using antibodies specific for TCRbeta-variable regions. Such (alphabeta)2TCR/CD3 or higher-order complexes were evident in T cells studied either ex vivo or after expansion in vitro. T cell activation is thought by many, but not all, to require TCR cross-linking by its antigen/major histocompatibility complex ligand. The implications of a multivalent (alphabeta)2TCR/CD3 complex stoichiometry for the ordered docking of specific antigen/major histocompatibility complex, CD4, or CD8 coreceptors and additional TCRs are discussed. 相似文献