4 research outputs found
Optical diagnostics of thermoplastic dielectric and semiconductor material in thin layers
No full textElectrographic and electrophotographic reserve materials with dirigible, structural, sensitometric, and operational characteristics
No full textImproving the resolution of interaction maps: A middleground between high-resolution complexes and genome-wide interactomes
No full textProtein-protein interactions are ubiquitous in Biology and therefore central to understand living organisms. In recent years, large-scale studies have been undertaken to describe, at least partially, protein-protein interaction maps or interactomes for a number of relevant organisms including human. Although the analysis of interaction networks is proving useful, current interactomes provide a blurry and granular picture of the molecular machinery, i.e. unless the structure of the protein complex is known the molecular details of the interaction are missing and sometime is even not possible to know if the interaction between the proteins is direct, i.e. physical interaction or part of functional, not necessary, direct association. Unfortunately, the determination of the structure of protein complexes cannot keep pace with the discovery of new protein-protein interactions resulting in a large, and increasing, gap between the number of complexes that are thought to exist and the number for which 3D structures are available. The aim of the thesis was to tackle this problem by implementing computational approaches to derive structural models of protein complexes and thus reduce this existing gap. Over the course of the thesis, a novel modelling algorithm to predict the structure of protein complexes, V-D2OCK, was implemented. This new algorithm combines structure-based prediction of protein binding sites by means of a novel algorithm developed over the course of the thesis: VORFFIP and M-VORFFIP, data-driven docking and energy minimization. This algorithm was used to improve the coverage and structural content of the human interactome compiled from different sources of interactomic data to ensure the most comprehensive interactome. Finally, the human interactome and structural models were compiled in a database, V-D2OCK DB, that offers an easy and user-friendly access to the human interactome including a bespoken graphical molecular viewer to facilitate the analysis of the structural models of protein complexes. Furthermore, new organisms, in addition to human, were included providing a useful resource for the study of all known interactomes
Machine learning and dengue forecasting: Comparing random forests and artificial neural networks for predicting dengue burden at national and sub-national scales in Colombia.
No full textThe robust estimate and forecast capability of random forests (RF) has been widely recognized, however this ensemble machine learning method has not been widely used in mosquito-borne disease forecasting. In this study, two sets of RF models were developed at the national (pooled department-level data) and department level in Colombia to predict weekly dengue cases for 12-weeks ahead. A pooled national model based on artificial neural networks (ANN) was also developed and used as a comparator to the RF models. The various predictors included historic dengue cases, satellite-derived estimates for vegetation, precipitation, and air temperature, as well as population counts, income inequality, and education. Our RF model trained on the pooled national data was more accurate for department-specific weekly dengue cases estimation compared to a local model trained only on the department's data. Additionally, the forecast errors of the national RF model were smaller to those of the national pooled ANN model and were increased with the forecast horizon increasing from one-week-ahead (mean absolute error, MAE: 9.32) to 12-weeks ahead (MAE: 24.56). There was considerable variation in the relative importance of predictors dependent on forecast horizon. The environmental and meteorological predictors were relatively important for short-term dengue forecast horizons while socio-demographic predictors were relevant for longer-term forecast horizons. This study demonstrates the potential of RF in dengue forecasting with a feasible approach of using a national pooled model to forecast at finer spatial scales. Furthermore, including sociodemographic predictors is likely to be helpful in capturing longer-term dengue trends
