Untitled

Funder

Organizational Unit

Publications

Stacked Denoising Autoencoders and Transfer Learning for Immunogold Particles Detection and Recognition

Publication . Sousa, Ricardo; Santos, Jorge M.; Silva, Luís M.; Alexandre, Luís; Esteves, Tiago; Rocha, Sara; Monjardino, Paulo; Sá, Joaquim Marques de; Figueiredo, Francisco; Quelhas, Pedro

In this paper we present a system for the detection of immunogold particles and a Transfer Learning (TL) framework for the recognition of these immunogold particles. Immunogold particles are part of a high-magnification method for the selective localization of biological molecules at the subcellular level only visible through Electron Microscopy. The number of immunogold particles in the cell walls allows the assessment of the differences in their compositions providing a tool to analise the quality of different plants. For its quantization one requires a laborious manual labeling (or annotation) of images containing hundreds of particles. The system that is proposed in this paper can leverage significantly the burden of this manual task. For particle detection we use a LoG filter coupled with a SDA. In order to improve the recognition, we also study the applicability of TL settings for immunogold recognition. TL reuses the learning model of a source problem on other datasets (target problems) containing particles of different sizes. The proposed system was developed to solve a particular problem on maize cells, namely to determine the composition of cell wall ingrowths in endosperm transfer cells. This novel dataset as well as the code for reproducing our experiments is made publicly available. We determined that the LoG detector alone attained more than 84\% of accuracy with the F-measure. Developing immunogold recognition with TL also provided superior performance when compared with the baseline models augmenting the accuracy rates by 10\%.

2017-12-07Preprint

Open access

ESBL and AmpC β-Lactamases in Clinical Strains of Escherichia coli from Serra da Estrela, Portugal

Publication . Oliveira, Cátia Marlene Meneses; Amador, Paula; Prudêncio, C.; Tomaz, C. T.; Ratado, Paulo; Fernandes, Ruben

Background and Objectives: Given the considerable spatial, temporal, and ecological factors, heterogeneity, which affects emergency response, persistence, and dissemination of genetic determinants that confer microorganisms their resistance to antibiotics, several authors claim that antibiotics' resistance must be perceived as an ecological problem. The aim of this study was to determine the prevalence of broad-spectrum bla genes, not only Extended-spectrum β-lactamases (ESBL) but also AmpC-types, in clinical strains of Escherichia coli isolated from Portugal (in the highest region of the country, Serra da Estrela) to disclose susceptibility profiles among different genotypes, and to compare the distribution of bla genes expressing broad-spectrum enzymes. Materials and Methods: Clinical strains of Escherichia coli presenting resistance to third generation (3G) cephalosporins and susceptibility to inhibition by clavulanic acid were studied by means of phenotypic and molecular profiling techniques for encoding β-lactamases genes. Results: Strains were mainly isolated from hospital populations (97%). Molecular analysis enabled the detection of 49 bla genes, in which 55% (27/49) were identified as blaOXA-1-like, 33% (16/49) as blaCTX-M-group-1, 10% (5/49) as blaTEM, and 2% (1/49) were identified as genes blaCIT (AmpC). Among all blaOXA-1-like detected, about 59% of strains expressed at least another bla gene. Co-production of β-lactamases was observed in 40% of strains, with the co-production of CTX-M group 1 and OXA-1-like occurring as the most frequent. Conclusions: This is the first study using microorganisms isolated from native people from the highest Portuguese mountain regions, showing an unprecedent high prevalence of genes blaOXA-1-like in this country.

2019Journal article

Open access