Machine Learning Using Neural Networks for Metabolomic Pathway Analyses.

Elucidating the mechanisms of metabolic pathways helps us understand the cascade of enzyme-catalyzed reactions that lead to the conversion of substances into final products. This has implications for predicting how newly synthesized compounds will affect a person's metabolism and, hence, the development of novel treatments to improve one's health. The study of metabolomic pathways, together with protein engineering, may also aid in the extraction, at a scale, of natural products to be used as drugs and drug precursors. Several approaches have been used to correlate protein annotations to metabolic pathways in order to derive pathways directly related to specific organisms. These could range from association rule-mining techniques to machine learning methods such as decision trees, naïve Bayes, logistic regression, and ensemble methods.In this chapter, we will be reviewing the use of machine learning for metabolic pathway analyses, with a step-by-step focus on the use of deep learning to predict the association of compounds (metabolites) to their respective metabolomic pathway classes. This prediction could help explain interactions of small molecules in organisms. Inspired by the work of Baranwal et al. (2019), we demonstrate how to build and train a deep learning neural network model to perform a multi-label prediction. We considered two different types of fingerprints as features (inputs to the model). The output of the model is the set of metabolic pathway classes (from the KEGG dataset) in which the input molecule participates. We will walk through the various steps of this process, including data collection, feature engineering, model selection, training, and evaluation. This model-building and evaluation process may be easily transferred to other domains of interest. All the source code used in this chapter is made publicly available at https://github.com/jp-um/machine_learning_for_metabolomic_pathway_analyses .

CTContour: An open-source Python pipeline for automatic contouring and calculation of mean SSDE along the abdomino-pelvic region for CT images; validation on fifteen systems.

PURPOSE: Calculation of the Size Specific Dose Estimate (SSDE) requires accurate delineation of the skin boundary of patient CT slices. The AAPM recommendation for SSDE evaluation at every CT slice is too time intensive for manual contouring, prohibiting real-time or bulk processing; an automated approach is therefore desirable. Previous automated delineation studies either did not fully disclose the steps of the algorithm or did not always manage to fully isolate the patient. The purpose of this study was to develop a validated, freely available, fast, vendor-independent open-source tool to automatically and accurately contour and calculate the SSDE for the abdomino-pelvic region for entire studies in real-time, including flagging of patient-truncated images. METHODS: The Python tool, CTContour, consists of a sequence of morphological steps and scales over multiple cores for speed. Tool validation was achieved on 700 randomly selected slices from abdominal and abdomino-pelvic studies from public datasets. Contouring accuracy was assessed visually by four medical physicists using a 1-5 Likert scale (5 indicating perfect contouring). Mean SSDE values were validated via manual calculation. RESULTS: Contour accuracy validation produced a score of four of five for 98.5 % of the images. A 300 slice exam was contoured and truncation flagged in 6.3 s on a six-core laptop. CONCLUSIONS: The algorithm was accurate even for complex clinical scenarios and when artefacts were present. Fast execution makes it possible to automate the calculation of SSDE in real time. The tool has been published on GitHub under the GNU-GPLv3 license.

Automated segmentation of microtomography imaging of Egyptian mummies.

Propagation Phase Contrast Synchrotron Microtomography (PPC-SRµCT) is the gold standard for non-invasive and non-destructive access to internal structures of archaeological remains. In this analysis, the virtual specimen needs to be segmented to separate different parts or materials, a process that normally requires considerable human effort. In the Automated SEgmentation of Microtomography Imaging (ASEMI) project, we developed a tool to automatically segment these volumetric images, using manually segmented samples to tune and train a machine learning model. For a set of four specimens of ancient Egyptian animal mummies we achieve an overall accuracy of 94-98% when compared with manually segmented slices, approaching the results of off-the-shelf commercial software using deep learning (97-99%) at much lower complexity. A qualitative analysis of the segmented output shows that our results are close in terms of usability to those from deep learning, justifying the use of these techniques.

Vision-Based Tunnel Lining Health Monitoring via Bi-Temporal Image Comparison and Decision-Level Fusion of Change Maps.

Tunnel structural health inspections are predominantly done through periodic visual observations, requiring humans to be physically present on-site, possibly exposing them to hazardous environments. These surveys are subjective (relying on the surveyor experience), time-consuming, and may demand operation shutdown. These issues can be mitigated through accurate automatic monitoring and inspection systems. In this work, we propose a remotely operated machine vision change detection application to improve the structural health monitoring of tunnels. The vision-based sensing system acquires the data from a rig of cameras hosted on a robotic platform that is driven parallel to the tunnel walls. These data are then pre-processed using image processing and deep learning techniques to reduce nuisance changes caused by light variations. Image fusion techniques are then applied to identify the changes occurring in the tunnel structure. Different pixel-based change detection approaches are used to generate temporal change maps. Decision-level fusion methods are then used to combine these change maps to obtain a more reliable detection of the changes that occur between surveys. A quantitative analysis of the results achieved shows that the proposed change detection system achieved a recall value of 81%, a precision value of 93% and an F1-score of 86.7%.

Machine learning techniques for protein function prediction.

Proteins play important roles in living organisms, and their function is directly linked with their structure. Due to the growing gap between the number of proteins being discovered and their functional characterization (in particular as a result of experimental limitations), reliable prediction of protein function through computational means has become crucial. This paper reviews the machine learning techniques used in the literature, following their evolution from simple algorithms such as logistic regression to more advanced methods like support vector machines and modern deep neural networks. Hyperparameter optimization methods adopted to boost prediction performance are presented. In parallel, the metamorphosis in the features used by these algorithms from classical physicochemical properties and amino acid composition, up to text-derived features from biomedical literature and learned feature representations using autoencoders, together with feature selection and dimensionality reduction techniques, are also reviewed. The success stories in the application of these techniques to both general and specific protein function prediction are discussed.