Ciencias Exactas y Ciencias de la Salud
Permanent URI for this collectionhttps://hdl.handle.net/11285/551039
Pertenecen a esta colección Tesis y Trabajos de grado de las Maestrías correspondientes a las Escuelas de Ingeniería y Ciencias así como a Medicina y Ciencias de la Salud.
Browse
Search Results
- Detection of epileptic seizures through brain waves analysis using Machine Learning algorithms.(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2021-11-17) Alvarado Elizalde, Cristian Yair; MARTINEZ LEDESMA, JUAN EMMANUEL; 200096; Martínez Ledesma, Juan Emmanuel; puemcuervo; Cuevas Díaz Durán, Raquel; Santos Díaz, Alejandro; Martínez Torteya, Antonio; School of Engineering and Sciences; Campus Estado de MéxicoElectroencephalogram(EEG) is an effective and non-invasive technique commonly used for monitoring brain activity. EEG readings are analyzed to determine changes in brain activity that may be useful for diagnosing neurological disorders and other seizure disorders. On the other hand, around 50 million people worldwide have epilepsy, making it one of the most common neurological diseases globally. The risk of premature death in people with epilepsy is up to three times higher than in the general population. Over the years, different researchers had been trying to detect seizures with different methods and with different approaches, but none algorithm has been fully implemented in the life of the people that have this disease, and for this reason, I developed a solution for this problem. The solution that I developed was to extract the information obtained by making a classification analysis using data acquired through the EEGs in a time-lapse of 1 second and once done, compare the results of the Machine Learning methods to find the best algorithms for solving the problem. The main objective of the algorithm is to find the most precise detection during epileptic seizures using public data, by extracting the temporal features from the electroencephalogram and with this learn the general structure of a seizure to make an effective detection in the less time possible.
- Pathway Identification and Drug Repurposing for Neurodegenerative Diseases: A Public RNA-seq Data Strategy(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2021-06-11) Marini Macouzet, Constanza; Martínez Ledesma, Juan Emmanuel; tolmquevedo; Aguirre Gamboa, Raúl; Treviño Alvarado, Víctor Manuel; Cuevas Díaz Durán, Raquel; School of Engineering and Sciences; Campus Estado de MéxicoThe thesis desertion detailed here intends to fulfill the requirements for the Master’s in Computer Science. Neurodegenerative human diseases are a significant problem due to their high prevalence. It is estimated that 1 to 2\% of people over the age of 60 years has Parkinson’s disease. The population with Alzheimer disease is approximately 24 million, and it is expected to double. In the US, 30,000 people are estimated to have Huntington Disease (HD). These diseases have no cure, they are fatal, and the quality of life of the patient is appalling. In this study, a comprehensive pipeline comprised of a preprocessing phase and an analytic phase of seven analyses was developed. The main objective was to interrogate public RNA-seq expression data in order to better understand the pathophysiology of the mentioned neurodegenerative diseases (NDs). The analytic phase includes identification of molecular subtypes of the disease, brain and immune cell types proportion estimation, differential gene expression analysis, bypass of copy number variation evaluation, functional analysis, drug repurposing, and comparison of the obtained results with those contained in genome-wise association studies. The interrogation provided key pathways of each disease, and routes that are shared among the NDs. Moreover, pivotal genes resulted in the comparison between diseases and tissues, and were paired with annotated and unannotated genome variants; for HD a possible blood biomarker was suggested. Finally, more than 30 known drugs were repurposed as a plausible treatment to these NDs.
- Association of gene expression signatures with genomic alterations and clinical outcomes(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2020-11-01) Ramos García, Axel Alejandro; MARTINEZ LEDESMA, JUAN EMMANUEL; 200096; Martínez Ledesma, Juan Emmanuel; puelquio, emipsanchez; Treviño Alvarado, Víctor Manuel; Cuevas Díaz Durán, Raquel; Aguirre Gamboa, Raúl; Escuela de Ingeniería y Ciencias; Campus MonterreyTechnological advances applied to molecular biology, have led this discipline to perform several and more complex experiments, which outcomes have been summarized within massive databases, provoking the emergence of new disciplines as well as innovative approaches to analyze this bunch of data. One of these disciplines is Bioinformatics, where high-throughput data have been utilized to understand some diseases, such as cancer, which has been studied in order to provide a better classification, diagnosis, and provide new possible treatments to this condition. Available data go, from whole-genome sequencing to tissue images, proteomic, and metabolomic, etc. In the case of gene expression profiles, one of the most utilized study approaches is the performance of single-gene analysis, a test which consists in the measurement of the level of expression gene by gene, carrying out a comparison between the case and control samples by a statistical method (t-test, Wilcoxon-rank-sum), to assign a p-value related to every gene, then by a threshold filter process, we will be able to identify significant genes, and finally, proceed to give a biological interpretation from obtained results. However, this approach presents some lacks, within which, we can mention: Due to the adjustment process, (necessary for the number of tests performed) can lead to information loss, labeling wrongly as false-negative some relevant genes. The use of arbitrary threshold values, provokes discoveries to be falsely positive if the values for higher values or false negatives for lower values. Modifications in biological processes are related to groups of genes, thus, measuring the variation of the expression level of these groups of genes will let us to give a better biological interpretation. These groups of genes have been identified and nowadays we can find them within several public databases, these collections of gene sets are known as gene-set, and they could be used to provide better insight when analyzing expression data. Thus, the purpose of this thesis was to find, if the score-gotten through single-sample gene set enrichment analysis from the bibliography, Hallmark, Oncogenic, CMAP Up, CMAP Down collections is relevant to perform cancer subtype-classification by unsupervised learning techniques (Hierarchical clustering), identify involved pathways in the gene mutation presence or absence. Finally, re- late this score with the survival probability, we were able to determine the life expectancy of people and candidate treatment drugs, based on the level of expression from the determined gene set, related to a specific biological process, chemical alteration, or aberration.
- Histopathological image classification using deep learning(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2020-11) Arredondo Padilla, Braulio; Martínez Ledesma, Juan Emmanuel; emipsanchez; Tamez Peña, José Gerardo; Santos Díaz, Alejandro; Martínez Torteya, Antonio; Escuela de Ingeniería y ciencias; Campus MonterreyThis thesis presents a study of digital pathology classification using and combining several techniques of machine learning and deep learning. Cancer is one of the most common causes of death around the world. One of the main complications of the disease is the prediction in the final stage. Nowadays there are many different studies to obtain a correct diagnosis on time. Some of these studies are tissue biopsies. These samples are analyzed by a pathologist, which must observe pixel by pixel a whole image of high dimensions to give a diagnostic of the disease, including stage and class. This activity takes weeks, even for experts, because usually several samples are extracted from a single patient. To speed up and facilitate this process, several models have been developed for digital pathology classification. With these models, it is easier to discard many patient slides than the traditional method, then, the main activity for a pathologist is to confirm a diagnosis with the most relevant or complicated sample. The downside of these models is that most of them are based on deep learning, a technique that is well known for its great performance, but also for its high requirements like graphic processors and memory resources. Consequently, we performed a complete analysis of several convolutional neural networks used in different ways to compare outcomes and efficiency. In addition, we include techniques such as recurrent neural networks and machine learning. Several models of deep learning and machine learning are presented as alternatives to convolutional neural networks, including 5 computer vision techniques. The main objective of our project is to perform a real alternative capable to achieve similar outcomes to deep learning with limited resources. The experiments were successful, including a real alternative for deep learning for the classification of 3 different types of cancer with an area under the curve higher than 90%.
- Siamese neural networks for few-shot birdsong classification(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2020-06) Rentería Aguilar, Sergio Santiago; Martínez Ledesma, Juan Emmanuel; lagdtorre/tolmquevedo; Taylor, Charles E.; Rascón Estebané, Caleb Antonio; Monroy Borja, Raúl; Escuela de Ingeniería y Ciencias; Campus Estado de MéxicoBird vocalizations have been the focus of a wide variety of interdisciplinary studies in bioacoustics and neuroethology since they serve as models of motor control, learning and auditory perception. Yet, researchers have only begun to shed light on the structure and function of birdsong. Hypotheses abound, but still there is little agreement as how songs should be analyzed. One of the main challenges has been to classify acoustic units (syllables) from birdsong recordings, a task requiring robust classification algorithms capable of generalizing to unseen instances and dealing with data scarcity. Systematically detecting changes in syllable repertoires can help biologists to understand the origin and evolution of birdsong. The process of learning good features to discriminate among numerous and different sound classes is computationally expensive. Moreover, it might be impossible to achieve acceptable performance in cases where training data is scarce and classes are unbalanced. To address this issue, we propose a few-shot learning task in which an algorithm must make predictions given only a few instances of each class. We compared the performance of different Siamese Neural Networks at metric learning over the set of Cassini’s Vireo syllables. Then, the network features were reused for the few-shot classification task. With this approach we overcame the limitations of data scarcity and class imbalance while achieving state-of-the-art performance.