UT Austin | Portugal Workshop on Nonlinear Mechanics and Applications in Life Sciences

This is a CoLab Workshop organized as an initiative of the UT Austin | Portugal Program to reinforce the Portuguese competences in Nonlinear Mechanics and in complex problems arising from applications to the mathematical modeling and simulations in the Life Sciences. The Workshop provides a place to exchange recent developments, discoveries and progresses in this challenging research field. The main goal is to bring together doctoral candidates, postdoctoral scientists and graduates interested in the field, giving them the opportunity to make scientific interactions and new connections with established experts in the interdisciplinary topics covered by the event. Another important goal of the Workshop is to promote collaboration between members of the different areas of the UT Austin | Portugal community.

Development of unsupervised learning methods with applications to life sciences data

Machine Learning makes computers capable of performing tasks typically requiring human intelligence. A domain where it is having a considerable impact is the life sciences, allowing to devise new biological analysis protocols, develop patients’ treatments efficiently and faster, and reduce healthcare costs. This Thesis work presents new Machine Learning methods and pipelines for the life sciences focusing on the unsupervised field. At a methodological level, two methods are presented. The first is an “Ab Initio Local Principal Path” and it is a revised and improved version of a pre-existing algorithm in the manifold learning realm. The second contribution is an improvement over the Import Vector Domain Description (one-class learning) through the Kullback-Leibler divergence. It hybridizes kernel methods to Deep Learning obtaining a scalable solution, an improved probabilistic model, and state-of-the-art performances. Both methods are tested through several experiments, with a central focus on their relevance in life sciences. Results show that they improve the performances achieved by their previous versions. At the applicative level, two pipelines are presented. The first one is for the analysis of RNA-Seq datasets, both transcriptomic and single-cell data, and is aimed at identifying genes that may be involved in biological processes (e.g., the transition of tissues from normal to cancer). In this project, an R package is released on CRAN to make the pipeline accessible to the bioinformatic Community through high-level APIs. The second pipeline is in the drug discovery domain and is useful for identifying druggable pockets, namely regions of a protein with a high probability of accepting a small molecule (a drug). Both these pipelines achieve remarkable results. Lastly, a detour application is developed to identify the strengths/limitations of the “Principal Path” algorithm by analyzing Convolutional Neural Networks induced vector spaces. This application is conducted in the music and visual arts domains.

Regularization meets GreenAI: a new framework for image reconstruction in life sciences applications

Ill-conditioned inverse problems frequently arise in life sciences, particularly in the context of image deblurring and medical image reconstruction. These problems have been addressed through iterative variational algorithms, which regularize the reconstruction by adding prior knowledge about the problem's solution. Despite the theoretical reliability of these methods, their practical utility is constrained by the time required to converge. Recently, the advent of neural networks allowed the development of reconstruction algorithms that can compute highly accurate solutions with minimal time demands. Regrettably, it is well-known that neural networks are sensitive to unexpected noise, and the quality of their reconstructions quickly deteriorates when the input is slightly perturbed. Modern efforts to address this challenge have led to the creation of massive neural network architectures, but this approach is unsustainable from both ecological and economic standpoints. The recently introduced GreenAI paradigm argues that developing sustainable neural network models is essential for practical applications. In this thesis, we aim to bridge the gap between theory and practice by introducing a novel framework that combines the reliability of model-based iterative algorithms with the speed and accuracy of end-to-end neural networks. Additionally, we demonstrate that our framework yields results comparable to state-of-the-art methods while using relatively small, sustainable models. In the first part of this thesis, we discuss the proposed framework from a theoretical perspective. We provide an extension of classical regularization theory, applicable in scenarios where neural networks are employed to solve inverse problems, and we show there exists a trade-off between accuracy and stability. Furthermore, we demonstrate the effectiveness of our methods in common life science-related scenarios. In the second part of the thesis, we initiate an exploration extending the proposed method into the probabilistic domain. We analyze some properties of deep generative models, revealing their potential applicability in addressing ill-posed inverse problems.

[Evaluation of risk factors for Mastitis-Metritis-Agalactia in pig farms in Switzerland].

Mastitis-Metritis-Agalactia (MMA), also known as postpartum dysgalactia syndrome (PPDS) is the most important disease complex in sows after birth. The present study compared 30 MMA problem herds (over 12% of farrowing sows affected) with 30 control farms (less than 10% of farrowing sows affected) to identify risk factors and treatment incidence. Important risk factors identified were in gilts the integration into the herd after the first farrowing, in gestating sows firm fecal consistency as well as in lactating sows soiled troughs, a low flow rate (<2 liters per minute) in drinking nipples and a high prevalence of lameness. The treatment incidence was also significantly different between the two groups. The MMA prevalence could be reduced through optimization of husbandry, feeding and management, which could essentially diminish the use of antibiotics.

[Risk factors for the oral use of antibiotics and animal treatment incidence of weaners in Switzerland].

In the present study, risk factors for the use of oral antibiotics in weaned piglets were collected on 112 pig farms by a personal questionaire. The most common indication for an antibiotic group therapy was diarrhoea, and the most frequently used antibiotic was Colistin. On average, 27.33 daily doses in the control farms and 387.21 daily doses in the problem farms per 1000 weaners were administered on a given day. The significant risk factors in the multivariate model were poor hygiene in the water supply of suckling piglets, less than two doses ofprestarter feed daily, lack of an all-in-and-all-out production system in weaners, no herd book performance data analysis, and less than two of the legally prescribed veterinary visits per year. Furthermore, the treatment incidence of weaners for oral antibiotics was calculated on the basis of the drug inventory. This study provides evidence that the use of oral antibiotics in weaners can be reduced by interventions in hygiene and management.

Analysis of behavioral changes in dairy cows associated with claw horn lesions

Detecting lame cows is important in improving animal welfare. Automated tools are potentially useful to enable identification and monitoring of lame cows. The goals of this study were to evaluate the suitability of various physiological and behavioral parameters to automatically detect lameness in dairy cows housed in a cubicle barn. Lame cows suffering from a claw horn lesion (sole ulcer or white line disease) of one claw of the same hind limb (n=32; group L) and 10 nonlame healthy cows (group C) were included in this study. Lying and standing behavior at night by tridimensional accelerometers, weight distribution between hind limbs by the 4-scale weighing platform, feeding behavior at night by the nose band sensor, and heart activity by the Polar device (Polar Electro Oy, Kempele, Finland) were assessed. Either the entire data set or parts of the data collected over a 48-h period were used for statistical analysis, depending upon the parameter in question. The standing time at night over 12 h and the limb weight ratio (LWR) were significantly higher in group C as compared with group L, whereas the lying time at night over 12 h, the mean limb difference (△weight), and the standard deviation (SD) of the weight applied on the limb taking less weight were significantly lower in group C as compared with group L. No significant difference was noted between the groups for the parameters of heart activity and feeding behavior at night. The locomotion score of cows in group L was positively correlated with the lying time and △weight, whereas it was negatively correlated with LWR and SD. The highest sensitivity (0.97) for lameness detection was found for the parameter SD [specificity of 0.80 and an area under the curve (AUC) of 0.84]. The highest specificity (0.90) for lameness detection was present for Δweight (sensitivity=0.78; AUC=0.88) and LWR (sensitivity=0.81; AUC=0.87). The model considering the data of SD together with lying time at night was the best predictor of cows being lame, accounting for 40% of the variation in the likelihood of a cow being lame (sensitivity=0.94; specificity=0.80; AUC=0.86). In conclusion, the data derived from the 4-scale-weighing platform, either alone or combined with the lying time at night over 12 h, represent the most valuable parameters for automated identification of lame cows suffering from a claw horn lesion of one individual hind limb.

Simultaneous Multiscale Polyaffine Registration by Incorporating Deformation Statistics

Locally affine (polyaffine) image registration methods capture intersubject non-linear deformations with a low number of parameters, while providing an intuitive interpretation for clinicians. Considering the mandible bone, anatomical shape differences can be found at different scales, e.g. left or right side, teeth, etc. Classically, sequential coarse to fine registration are used to handle multiscale deformations, instead we propose a simultaneous optimization of all scales. To avoid local minima we incorporate a prior on the polyaffine transformations. This kind of groupwise registration approach is natural in a polyaffine context, if we assume one configuration of regions that describes an entire group of images, with varying transformations for each region. In this paper, we reformulate polyaffine deformations in a generative statistical model, which enables us to incorporate deformation statistics as a prior in a Bayesian setting. We find optimal transformations by optimizing the maximum a posteriori probability. We assume that the polyaffine transformations follow a normal distribution with mean and concentration matrix. Parameters of the prior are estimated from an initial coarse to fine registration. Knowing the region structure, we develop a blockwise pseudoinverse to obtain the concentration matrix. To our knowledge, we are the first to introduce simultaneous multiscale optimization through groupwise polyaffine registration. We show results on 42 mandible CT images.

The Psychological Outcome of Religious Coping with Stressful Life Events in a Swiss Sample of Church Attendees

Background: Recent research suggested thatreligious coping, based on dispositional religiousness and spirituality (R/S), is an important modulating factor in the process of dealing with adversity. In contrast to the United States, the effect of R/S on psychological adjustment to stress is a widely unexplored area in Europe. Methods: We examined a Swiss sample of 328 church attendees in the aftermath of stressful life events to explore associations of positive or negative religious coping with the psychological outcome. Applying a cross-sectional design, we used Huber’s Centrality Scale to specify religiousness and Pargament’s measure of religious coping (RCOPE) for the assessment of positive and negative religious coping. Depressive symptoms and anxiety as outcome variables were examined by the Brief Symptom Inventory. The Stress-Related Growth Scale and the Marburg questionnaire for the assessment of well-being were used to assess positive outcome aspects. We conducted Mann-Whitney tests for group comparisons and cumulative logit analysis for the assessmentof associations of religious coping with our outcome variables. Results: Both forms of religious coping were positively associated with stress-related growth (p < 0.01). However, negative religious coping additionally reduced well-being (p = 0.05, β = 0.52, 95% CI = 0.27–0.99) and increased anxiety (p = 0.02, β = 1.94, 95% CI = 1.10–3.39) and depressive symptoms (p = 0.01, β = 2.27, 95% CI = 1.27–4.06). Conclusions: The effects of religious coping on the psychological adjustment to stressful life events seem relevant. These findings should be confirmed in prospective studies.

One Approach to Drawing Tasks Statistics

Report published in the Proceedings of the National Conference on "Education and Research in the Information Society", Plovdiv, May, 2014

L'aprenentatge basat en problemes en models híbrids de docència-aprenentatge

La Facultat de Ciències de la Salut i de la Vida ha utilitzat des de 2004 la metodologia d'aprenentatge basat en problemes (en endavant ABP) com a mètode docent en els seus estudis de Biologia. En aquest període hem après algunes de les claus de l'aplicació del mètode en els nostres estudis. En primer lloc, cal disposar d'elements formatius que afavoreixin la formació dels tutors que participin en el projecte. Per assolir aquest objectiu hem dissenyat un portal on els nostres professors poden disposar de materials útils per a la seva activitat, així com de documents que permetin entendre millor el que suposa l'ABP. En segon lloc, el projecte tenia l'objectiu de dissenyar i avaluar activitats que permetessin integrar les pràctiques de laboratori en la lògica de la resolució de problemes pròpia de l'ABP. En aquest sentit vam dissenyar dues activitats en el tercer curs de la llicenciatura que anomenaren aprenentatge basat en el laboratori (ABL). Per aquest motiu es van dissenyar problemes que tinguessin una primera part de resolució a l'aula en grup de tutoria i una segona que obligués els estudiants a realitzar experiments de laboratori dirigits a entendre i resoldre les qüestions plantejades al grup de tutoria. L'ABL-1 fou un projecte de biologia cel·lular i destinat a aprofundir en els mecanismes implicats en els fenòmens de diferenciació dels miòcits. L'ABL-2 era un projecte conjunt dels professors de Fisiologia vegetal, Bioestadística i Microbiologia. En aquest cas es desitjava que els estudiants plantegessin la resolució a un problema que suposava la manipulació genètica de cèl·lules vegetals per fer possible que produïssin una substància específica, l'escopolamina. Finalment els estudiants havien d'escriure un article original com a projecte final de cada ABL. Els resultats dels dos anys d'experimentació han esta altament satisfactoris, d'acord amb les enquestes completades per alumnes i professors.

Effective diagnosis of Alzheimer's disease by means of large margin-based methodology.

BACKGROUND Functional brain images such as Single-Photon Emission Computed Tomography (SPECT) and Positron Emission Tomography (PET) have been widely used to guide the clinicians in the Alzheimer's Disease (AD) diagnosis. However, the subjectivity involved in their evaluation has favoured the development of Computer Aided Diagnosis (CAD) Systems. METHODS It is proposed a novel combination of feature extraction techniques to improve the diagnosis of AD. Firstly, Regions of Interest (ROIs) are selected by means of a t-test carried out on 3D Normalised Mean Square Error (NMSE) features restricted to be located within a predefined brain activation mask. In order to address the small sample-size problem, the dimension of the feature space was further reduced by: Large Margin Nearest Neighbours using a rectangular matrix (LMNN-RECT), Principal Component Analysis (PCA) or Partial Least Squares (PLS) (the two latter also analysed with a LMNN transformation). Regarding the classifiers, kernel Support Vector Machines (SVMs) and LMNN using Euclidean, Mahalanobis and Energy-based metrics were compared. RESULTS Several experiments were conducted in order to evaluate the proposed LMNN-based feature extraction algorithms and its benefits as: i) linear transformation of the PLS or PCA reduced data, ii) feature reduction technique, and iii) classifier (with Euclidean, Mahalanobis or Energy-based methodology). The system was evaluated by means of k-fold cross-validation yielding accuracy, sensitivity and specificity values of 92.78%, 91.07% and 95.12% (for SPECT) and 90.67%, 88% and 93.33% (for PET), respectively, when a NMSE-PLS-LMNN feature extraction method was used in combination with a SVM classifier, thus outperforming recently reported baseline methods. CONCLUSIONS All the proposed methods turned out to be a valid solution for the presented problem. One of the advances is the robustness of the LMNN algorithm that not only provides higher separation rate between the classes but it also makes (in combination with NMSE and PLS) this rate variation more stable. In addition, their generalization ability is another advance since several experiments were performed on two image modalities (SPECT and PET).

Consistency from the perspective of an experimental systems approach to the sciences and their epistemic objects

It is generally accepted that the development of the modern sciences is rooted in experiment. Yet for a long time, experimentation did not occupy a prominent role, neither in philosophy nor in history of science. With the 'practical turn' in studying the sciences and their history, this has begun to change. This paper is concerned with systems and cultures of experimentation and the consistencies that are generated within such systems and cultures. The first part of the paper exposes the forms of historical and structural coherence that characterize the experimental exploration of epistemic objects. In the second part, a particular experimental culture in the life sciences is briefly described as an example. A survey will be given of what it means and what it takes to analyze biological functions in the test tube.