High-Content Analysis of Breast Cancer Using Single-Cell Deep Transfer Learning

High-content analysis has revolutionized cancer drug discovery by identifying substances that alter the phenotype of a cell, which prevents tumor growth and metastasis. The high-resolution biofluorescence images from assays allow precise quantitative measures enabling the distinction of small molecules of a host cell from a tumor. In this work, we are particularly interested in the application of deep neural networks (DNNs), a cutting-edge machine learning method, to the classification of compounds in chemical mechanisms of action (MOAs). Compound classification has been performed using image-based profiling methods sometimes combined with feature reduction methods such as principal component analysis or factor analysis. In this article, we map the input features of each cell to a particular MOA class without using any treatment-level profiles or feature reduction methods. To the best of our knowledge, this is the first application of DNN in this domain, leveraging single-cell information. Furthermore, we use deep transfer learning (DTL) to alleviate the intensive and computational demanding effort of searching the huge parameter's space of a DNN. Results show that using this approach, we obtain a 30% speedup and a 2% accuracy improvement.

Virtual communities of practice: investigating motivations and constraints in the processes of knowledge creation and transfer

With accelerated market volatility, faster response times and increased globalization, business environments are going through a major transformation and firms have intensified their search for strategies which can give them competitive advantage. This requires that companies continuously innovate, to think of new ideas that can be transformed or implemented as products, processes or services, generating value for the firm. Innovative solutions and processes are usually developed by a group of people, working together. A grouping of people that share and create new knowledge can be considered as a Community of Practice (CoP). CoP’s are places which provide a sound basis for organizational learning and encourage knowledge creation and acquisition. Virtual Communities of Practice (VCoP's) can perform a central role in promoting communication and collaboration between members who are dispersed in both time and space. Nevertheless, it is known that not all CoP's and VCoP's share the same levels of performance or produce the same results. This means that there are factors that enable or constrain the process of knowledge creation. With this in mind, we developed a case study in order to identify both the motivations and the constraints that members of an organization experience when taking part in the knowledge creating processes of VCoP's. Results show that organizational culture and professional and personal development play an important role in these processes. No interviewee referred to direct financial rewards as a motivation factor for participation in VCoPs. Most identified the difficulty in aligning objectives established by the management with justification for the time spent in the VCoP. The interviewees also said that technology is not a constraint.

Virtual communities of practice: examining the motivations and constraints in their knowledge creation and knowledge transfer preocesses

Paper accepted for the OKLC 2009 - International Conference on Organizational Learning, Knowledge and Capabilities (26-28th, April 2009, Amsterdam, the Netherlands).

Desenvolvimento de metodologias analíticas para a determinação de resíduos de pesticidas em Vinhos do Douro e Vinhos do Porto

A produção de vinho é uma actividade de elevada importância ao nível económico, ambiental e social. Como tal, a protecção da vinha e/ou da uva através da utilização de pesticidas assume um papel fundamental nesta actividade, permitindo um aumento no rendimento da produção. No entanto, a transferência destes compostos da uva para o vinho é inevitável, ocorrendo, na maioria dos casos, em quantidades reduzidas. Apesar de, geralmente, a quantidade de pesticidas que pode ser transferida para o vinho não apresentar problemas relevantes para a saúde pública, o desenvolvimento de metodologias que permitam garantir um controlo rigoroso da qualidade do vinho é de elevada importância. O controlo deste produto, assim como de qualquer produto alimentar, não deve ser de certa forma virtual, levando a uma desconfiança crescente do consumidor. Ao longo deste trabalho foram desenvolvidos dois métodos para a determinação de pesticidas em diferentes tipos de vinhos, por SPME-GC-MS/MS. O primeiro método desenvolvido visa a determinação de 8 pesticidas organofosforados e o segundo método permite a determinação de 7 pesticidas, 4 organoclorados e 3 dicarboximidas. Foram estudadas quatro matrizes diferentes: vinho branco do Douro, vinho tinto do Douro, vinho branco do Porto e vinho tinto do Porto. As metodologias adoptadas permitiram a obtenção de bons resultados, apesar das condições adoptadas não serem as mais vantajosas para alguns dos compostos estudados. Dada a complexidade das matrizes, a detecção por espectrometria de massa mostrou ser fundamental para a identificação inequívoca de cada um dos pesticidas. Verificaram-se, de um modo geral, bons resultados ao nível da linearidade, para as gamas de concentrações escolhidas, para os dois métodos. Obtiveram-se também bons resultados para os limites de detecção e quantificação, cujos valores se situam abaixo dos limites máximos de resíduos para as uvas, impostos pela regulamentação europeia, para a maioria dos compostos. No caso dos organofosforados, os limites de detecção variam entre 0,05 μg/L e 13,00 μg/L para os pesticidas clorpirifos-metilo e metidatião, respectivamente. Os limites de quantificação variam entre 0,18 μg/L e 43,32 μg/L, também para os pesticidas clorpirifos-metilo e metidatião. No que se refere ao método para os pesticidas organoclorados e dicarboximidas, foram calculados vários limites de detecção para cada composto, em cada matriz estudada, a partir das rectas de calibração diárias, resultando num conjunto de valores com variações significativas entre cada um, para o mesmo pesticida e na mesma matriz. No entanto, apenas os limites de quantificação foram validados, recorrendo-se aos estudos de repetibilidade e precisão intermédia. Obtiveram-se os seguintes limites de quantificação: captana 52,10 μg/L; clortalonil 20,95 μg/L; dicofol 4,37 μg/L; folpete 93,60 μg/L; iprodiona 274,70 μg/L; procimidona 76,04 μg/L e vinclozolina 10,03 μg/L. Os pesticidas metidatião e captana apresentam-se como os compostos mais problemáticos, uma vez que os limites de quantificação obtidos não permitem a garantia do cumprimento dos limites máximos de resíduos regulamentados. Demonstra-se também que, apesar do efeito de matriz ser significativo na determinação de todos os compostos analisados, o mesmo pode ser atenuado, procedendo-se às calibrações nas respectivas matrizes. Observaram-se variações significativas ao nível da resposta do equipamento ao longo do tempo, quer pela alteração das condições operatórias, quer pela decomposição de alguns compostos ao longo do tempo. Este efeito, evidenciado nos estudos das precisões intermédias e repetibilidades, levou à adopção de um método de validação diferente para a determinação de pesticidas organoclorados e dicarboximidas, recorrendo-se então a calibrações diárias e sequências de trabalho mais pequenas.

The invisible factors that can make the difference. Language management and knowledge transfer in multinational sites

The knowledge-based society we live in has stressed the importance of human capital and brought talent to the top of most wanted skills, especially to companies who want to succeed in turbulent environments worldwide. In fact, streams, sequences of decisions and resource commitments characterize the day-to-day of multinational companies (MNCs). Such decision-making activities encompass major strategic moves like internationalization and new market entries or diversification and acquisitions. In most companies, these strategic decisions are extensively discussed and debated and are generally framed, formulated, and articulated in specialized language often developed by the best minds in the company. Yet the language used in such deliberations, in detailing and enacting the implementation strategy is usually taken for granted and receives little if any explicit attention (Brannen & Doz, 2012) an can still be a “forgotten factor” (Marschan et al. 1997). Literature on language management and international business refers to lack of awareness of business managers of the impact that language can have not only in communication effectiveness but especially in knowledge transfer and knowledge management in business environments. In the context of MNCs, management is, for many different reasons, more complex and demanding than that of a national company, mainly because of diversity factors inherent to internationalization, namely geographical and cultural spaces, i.e, varied mindsets. Moreover, the way of functioning, and managing language, of the MNC depends on its vision, its values and its internationalization model, i.e on in the way the MNE adapts to and controls the new markets, which can vary essentially from a more ethnocentric to a more pluricentric focus. Regardless of the internationalization model followed by the MNC, communication between different business units is essential to achieve unity in diversity and business sustainability. For the business flow and prosperity, inter-subsidiary, intra-company and company-client (customers, suppliers, governments, municipalities, etc..) communication must work in various directions and levels of the organization. If not well managed, this diversity can be a barrier to global coordination and create turbulent environments, even if a good technological support is available (Feely et al., 2002: 4). According to Marchan-Piekkari (1999) the tongue can be both (i) a barrier, (ii) a facilitator and (iii) a source of power. Moreover, the lack of preparation for the barriers of linguistic diversity can lead to various costs, including negotiations’ failure and failure on internationalization.. On the other hand, communication and language fluency is not just a message transfer procedure, but above all a knowledge transfer process, which requires extra-linguistic skills (persuasion, assertiveness …) in order to promote credibility of both parties. For this reason, MNCs need a common code to communicate and trade information inside and outside the company, which will require one or more strategies, in order to overcome possible barriers and organization distortions.

Heat transfer coefficients from Newtonian and non-Newtonian fluids flowing in laminar regime in a helical coil

This study aimed to carry out experimental work to obtain, for Newtonian and non-Newtonian fluids, heat transfer coefficients, at constant wall temperature as boundary condition, in fully developed laminar flow inside a helical coil. The Newtonian fluids studied were aqueous solutions of glycerol, 25%, 36%, 43%, 59% and 78% (w/w) and the non-Newtonian fluids aqueous solutions of carboxymethylcellulose (CMC), a polymer, with concentrations 0.1%, 0.2%, 0.3%, 0.4% and 0.6% (w/w) and aqueous solutions of xanthan gum (XG), another polymer, with concentrations 0.1% and 0.2% (w/w). According to the rheological study performed, the polymer solutions had shear thinning behavior and different values of elasticity. The helical coil used has internal diameter, curvature ratio, length and pitch, respectively: 0.004575 m, 0.0263, 5.0 m and 11.34 mm. The Nusselt numbers for the CMC solutions are, on average, slightly higher than those for Newtonian fluids, for identical Prandtl and generalized Dean numbers. As outcome, the viscous component of the shear thinning polymer tends to potentiate the mixing effect of the Dean cells. The Nusselt numbers of the XG solutions are significant lower than those of the Newtonian solutions, for identical Prandtl and generalized Dean numbers. Therefore, the elastic component of the polymer tends to diminish the mixing effect of the Dean cells. A global correlation, for Nusselt number as a function of Péclet, generalized Dean and Weissenberg numbers for all Newtonian and non-Newtonian solutions studied, is presented.

Determinação do teor de compostos aromáticos policíclicos de óleos minerais isolantes para transformadores

Mestrado em Engenharia Química - Ramo Optimização Energética na Indústria Química

Finite Element Analysis (FEA) applied to heat transfer optimization process of pultrusion die systems

The aim of this study is to optimize the heat flow through the pultrusion die assembly system on the manufacturing process of a specific glass-fiber reinforced polymer (GFRP) pultrusion profile. The control of heat flow and its distribution through whole die assembly system is of vital importance in optimizing the actual GFRP pultrusion process. Through mathematical modeling of heating-die process, by means of Finite Element Analysis (FEA) program, an optimum heater selection, die position and temperature control was achieved. The thermal environment within the die was critically modeled relative not only to the applied heat sources, but also to the conductive and convective losses, as well as the thermal contribution arising from the exothermic reaction of resin matrix as it cures or polymerizes from the liquid to solid condition. Numerical simulation was validated with basis on thermographic measurements carried out on key points along the die during pultrusion process.