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.

RuSe Electrocatalysts and Single Wall Carbon Nanohorns Supports for the Oxygen Reduction Reaction

Selenium modified ruthenium electrocatalysts supported on carbon black were synthesized using NaBH4 reduction of the metal precursor. Prepared Ru/C electrocatalysts showed high dispersion and very small averaged particle size. These Ru/C electrocatalysts were subsequently modified with Se following two procedures: (a) preformed Ru/carbon catalyst was mixed with SeO2 in xylene and reduced in H2 and (b) Ru metal precursor was mixed with SeO2 followed by reduction with NaBH4. The XRD patterns indicate that a pyrite-type structure was obtained at higher annealing temperatures, regardless of the Ru:Se molar ratio used in the preparation step. A pyrite-type structure also emerged in samples that were not calcined; however, in this case, the pyrite-type structure was only prominent for samples with higher Ru:Se ratios. The characterization of the RuSe/C electrocatalysts suggested that the Se in noncalcined samples was present mainly as an amorphous skin. Preliminary study of activity toward oxygen reduction reaction (ORR) using electrocatalysts with a Ru:Se ratio of 1:0.7 indicated that annealing after modification with Se had a detrimental effect on their activity. This result could be related to the increased particle size of crystalline RuSe2 in heat-treated samples. Higher activity of not annealed RuSe/C catalysts could also be a result of the structure containing amorphous Se skin on the Ru crystal. The electrode obtained using not calcined RuSe showed a very promising performance with a slightly lower activity and higher overpotential in comparison with a commercial Pt/C electrode. Single wall carbon nanohorns (SWNH) were considered for application as ORR electrocatalysts' supports. The characterization of SWNH was carried out regarding their tolerance toward strong catalyzed corrosion conditions. Tests indicated that SWNH have a three times higher electrochemical surface area (ESA) loss than carbon black or Pt commercial electrodes.