This paper presents a novel modelling approach for complex nonlinear dynamic of the neural signal conduction along the myelinated or unmyelinated axons. Normally, this problem is described by the partial differential equation (PDE) combing cable equation, however the solution of the PDE approach is difficult to obtain and the interaction phenomena in nerve tissue is ignored. Based on radial basis function neural network (RBFNN), the membrane potential conduction can be restated by the dynamic of the weight vector while the shortcomings of the PDE approach can be fixed. Moreover, the neural signal prediction, the stimulation signal design and interaction characterization are further investigated using the presented model.
{"title":"RBFNN-based Modelling and Analysis for the Signal Reconstruction of Peripheral Nerve Tissue","authors":"Qichun Zhang, F. Sepulveda","doi":"10.1145/3107411.3107478","DOIUrl":"https://doi.org/10.1145/3107411.3107478","url":null,"abstract":"This paper presents a novel modelling approach for complex nonlinear dynamic of the neural signal conduction along the myelinated or unmyelinated axons. Normally, this problem is described by the partial differential equation (PDE) combing cable equation, however the solution of the PDE approach is difficult to obtain and the interaction phenomena in nerve tissue is ignored. Based on radial basis function neural network (RBFNN), the membrane potential conduction can be restated by the dynamic of the weight vector while the shortcomings of the PDE approach can be fixed. Moreover, the neural signal prediction, the stimulation signal design and interaction characterization are further investigated using the presented model.","PeriodicalId":246388,"journal":{"name":"Proceedings of the 8th ACM International Conference on Bioinformatics, Computational Biology,and Health Informatics","volume":"79 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116344744","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Common methods for the functional inference of genomic data, such as Gene Sent Enrichment Analysis (GSEA) and Over Representation Analysis (ORA), often discard the interactions between the biomolecular entities. Recent studies have explored this issue through a variety of techniques and show that using evidence from the interactions produces a more relevant and insightful inference. In this article, we introduce a method, referred to as Causal Disturbance (Cdist), for enrichment analysis. Cdist utilizes the underlying graph of pathways in combination with experimental data to detect the pathway dysregulations. To test our methodology, we utilized a public microarray data from colorectal cancer. We show that Cdist identifies the dysregulated pathways of colorectal cancer that are not detectable by other conventional methods. Some of the detected pathways by Cdist, such as apoptosis and Ras signaling, are critical for their roles in cancer. We conclude that our method facilitates a more informative inference of the disease data by incorporating the topological features of the pathway graphs. Using these features will help to detect the pathway dysregulations that are not observable through conventional models.
{"title":"Use of Structural Properties of Underlying Graphs in Pathway Enrichment Analysis of Genomic Data","authors":"Pourya Naderi Yeganeh, M. Mostafavi","doi":"10.1145/3107411.3107488","DOIUrl":"https://doi.org/10.1145/3107411.3107488","url":null,"abstract":"Common methods for the functional inference of genomic data, such as Gene Sent Enrichment Analysis (GSEA) and Over Representation Analysis (ORA), often discard the interactions between the biomolecular entities. Recent studies have explored this issue through a variety of techniques and show that using evidence from the interactions produces a more relevant and insightful inference. In this article, we introduce a method, referred to as Causal Disturbance (Cdist), for enrichment analysis. Cdist utilizes the underlying graph of pathways in combination with experimental data to detect the pathway dysregulations. To test our methodology, we utilized a public microarray data from colorectal cancer. We show that Cdist identifies the dysregulated pathways of colorectal cancer that are not detectable by other conventional methods. Some of the detected pathways by Cdist, such as apoptosis and Ras signaling, are critical for their roles in cancer. We conclude that our method facilitates a more informative inference of the disease data by incorporating the topological features of the pathway graphs. Using these features will help to detect the pathway dysregulations that are not observable through conventional models.","PeriodicalId":246388,"journal":{"name":"Proceedings of the 8th ACM International Conference on Bioinformatics, Computational Biology,and Health Informatics","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127952494","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
To analyze large DNA data sets, we hypothesized that the organization of repeated bases within DNA follows rules similar to Cellular Automata (CA). These sequences could be defined as regular domains. By considering DNA strings as a finite one-dimensional cell automated, consisting of a finite (numerable) set of cells spatially aligned on a straight line and adopting a color code that transforms the DNA bases (A, C, T, G) in numbers, we analyzed DNA strings in the approach of computational mechanics. In this approach, a regular domain is a space-time region consisting of sequences in the same regular language (the particular rule of system evolution, which gives rise to a formal language) that creates patterns computationally homogeneous and simple to describe. We discovered that regular domain exists. Results revealed the exact number of strings of given lengths, establishing their limit in length, their precise localizations in all the human chromosomes and their complex numerical organization. Furthermore, the distribution of these domains is not at random, nor chaotic neither probabilistic, but there are numeric attractors around which the number of these domains are distributed. This leads us to think that all these domains within the DNA are connected to each other and cannot be casually distributed, but they follow some combinatorics rules.
{"title":"Discovery of Regular Domains in Large DNA Data Sets","authors":"F. Bertacchini, E. Bilotta, Pietro S. Pantano","doi":"10.1145/3107411.3110419","DOIUrl":"https://doi.org/10.1145/3107411.3110419","url":null,"abstract":"To analyze large DNA data sets, we hypothesized that the organization of repeated bases within DNA follows rules similar to Cellular Automata (CA). These sequences could be defined as regular domains. By considering DNA strings as a finite one-dimensional cell automated, consisting of a finite (numerable) set of cells spatially aligned on a straight line and adopting a color code that transforms the DNA bases (A, C, T, G) in numbers, we analyzed DNA strings in the approach of computational mechanics. In this approach, a regular domain is a space-time region consisting of sequences in the same regular language (the particular rule of system evolution, which gives rise to a formal language) that creates patterns computationally homogeneous and simple to describe. We discovered that regular domain exists. Results revealed the exact number of strings of given lengths, establishing their limit in length, their precise localizations in all the human chromosomes and their complex numerical organization. Furthermore, the distribution of these domains is not at random, nor chaotic neither probabilistic, but there are numeric attractors around which the number of these domains are distributed. This leads us to think that all these domains within the DNA are connected to each other and cannot be casually distributed, but they follow some combinatorics rules.","PeriodicalId":246388,"journal":{"name":"Proceedings of the 8th ACM International Conference on Bioinformatics, Computational Biology,and Health Informatics","volume":"198 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132474245","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Diabetes afflicts an estimated over 400 million people worldwide. People with diabetes are at the risk of a wide range of devastating complications including diabetic peripheral neuropathy, which is commonly referred to as the "diabetic foot" and most often affects the lower extremities (i.e., leg and foot) and can lead to amputations. In this paper, we present a computer aided diagnostic system for diabetic foot. At the core of our system is an improved thermoregulation model that characterizes the thermal recovery process of the extremities of the body (e.g., foot) after a cold stress. The model consists of a series of differential equations which is developed based on physiological characterizations and yet also exhibits analytical solutions. The model has been shown to be accurate and robust. Based on the new thermal regulation model, we have developed a 2D Bayesian classifier. We have applied the classifier to a cohort of 49 subjects (35 with no diabetic peripheral neuropathy and 14 with diabetic peripheral neuropathy). The classifier can accurately diagnose 93% of the subjects with diabetic peripheral neuropathy with a false positive rate of only 6%. This significantly outperforms current clinical diagnostic methods which may miss 61% of the patients with diabetic peripheral neuropathy.
{"title":"A Physiological Thermal Regulation Model with Application to the Diagnosis of Diabetic Peripheral Neuropathy","authors":"V. Chekh, P. Soliz, M. Burge, S. Luan","doi":"10.1145/3107411.3107512","DOIUrl":"https://doi.org/10.1145/3107411.3107512","url":null,"abstract":"Diabetes afflicts an estimated over 400 million people worldwide. People with diabetes are at the risk of a wide range of devastating complications including diabetic peripheral neuropathy, which is commonly referred to as the \"diabetic foot\" and most often affects the lower extremities (i.e., leg and foot) and can lead to amputations. In this paper, we present a computer aided diagnostic system for diabetic foot. At the core of our system is an improved thermoregulation model that characterizes the thermal recovery process of the extremities of the body (e.g., foot) after a cold stress. The model consists of a series of differential equations which is developed based on physiological characterizations and yet also exhibits analytical solutions. The model has been shown to be accurate and robust. Based on the new thermal regulation model, we have developed a 2D Bayesian classifier. We have applied the classifier to a cohort of 49 subjects (35 with no diabetic peripheral neuropathy and 14 with diabetic peripheral neuropathy). The classifier can accurately diagnose 93% of the subjects with diabetic peripheral neuropathy with a false positive rate of only 6%. This significantly outperforms current clinical diagnostic methods which may miss 61% of the patients with diabetic peripheral neuropathy.","PeriodicalId":246388,"journal":{"name":"Proceedings of the 8th ACM International Conference on Bioinformatics, Computational Biology,and Health Informatics","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131077346","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Jingyun Choi, Yejin Kim, Hun‐Sung Kim, I. Choi, Hwanjo Yu
To automatically extract medical concepts from raw electronic health records (EHRs), several applications based on machine learning techniques have been proposed. Among the various techniques, tensor factorization methods have attracted considerable attention because tensor representations can capture interactions among high-dimensional EHRs. Most of the existing tensor factorization methods for computational phenotyping are only designed to derive individual phenotypes that approximate the original data. However, deriving grouped phenotypes is desirable because patients form natural groups of interest (i.e., efficacy of treatment and disease categories). In this paper, we propose Supervised Non-negative Tensor Factorization with Multinomial Logistic Regression (SNTFL) to derive grouped phenotypes that are discriminative. We define a discriminative constraint to derive grouped phenotypes and jointly optimize a multinomial logistic regression during the tensor factorization process. Our case study on a hyperlipidemia dataset demonstrates that our proposed method obtains better discrimination on patient groups compared to the baselines and successfully discovers meaningful patient subgroups.
{"title":"Tensor-Factorization-Based Phenotyping using Group Information: Case Study on the Efficacy of Statins","authors":"Jingyun Choi, Yejin Kim, Hun‐Sung Kim, I. Choi, Hwanjo Yu","doi":"10.1145/3107411.3107423","DOIUrl":"https://doi.org/10.1145/3107411.3107423","url":null,"abstract":"To automatically extract medical concepts from raw electronic health records (EHRs), several applications based on machine learning techniques have been proposed. Among the various techniques, tensor factorization methods have attracted considerable attention because tensor representations can capture interactions among high-dimensional EHRs. Most of the existing tensor factorization methods for computational phenotyping are only designed to derive individual phenotypes that approximate the original data. However, deriving grouped phenotypes is desirable because patients form natural groups of interest (i.e., efficacy of treatment and disease categories). In this paper, we propose Supervised Non-negative Tensor Factorization with Multinomial Logistic Regression (SNTFL) to derive grouped phenotypes that are discriminative. We define a discriminative constraint to derive grouped phenotypes and jointly optimize a multinomial logistic regression during the tensor factorization process. Our case study on a hyperlipidemia dataset demonstrates that our proposed method obtains better discrimination on patient groups compared to the baselines and successfully discovers meaningful patient subgroups.","PeriodicalId":246388,"journal":{"name":"Proceedings of the 8th ACM International Conference on Bioinformatics, Computational Biology,and Health Informatics","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132173437","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Richard Platania, Shayan Shams, Seungwon Yang, Jian Zhang, Kisung Lee, Seung-Jong Park
Detection of suspicious regions in mammogram images and the subsequent diagnosis of these regions remains a challenging problem in the medical world. There still exists an alarming rate of misdiagnosis of breast cancer. This results in both over treatment through incorrect positive diagnosis of cancer and under treatment through overlooked cancerous masses. Convolutional neural networks have shown strong applicability to various image datasets, enabling detailed features to be learned from the data and, as a result, the ability to classify these images at extremely low error rates. In order to overcome the difficulty in diagnosing breast cancer from mammogram images, we propose our framework for automated breast cancer detection and diagnosis, called BC-DROID, which provides automated region of interest detection and diagnosis using convolutional neural networks. BC-DROID first pretrains based on physician-defined regions of interest in mammogram images. It then trains based on the full mammogram image. The resulting network is able to detect and classify regions of interest as cancerous or benign in one step. We demonstrate the accuracy of our framework's ability to both locate the regions of interest as well as diagnose them. Our framework achieves a detection accuracy of up to 90% and a classification accuracy of 93.5% (AUC of 92.315%). To the best of our knowledge, this is the first work enabling both automated detection and diagnosis of these areas in one step from full mammogram images. Using our framework's website, a user can upload a single mammogram image, visualize suspicious regions, and receive the automated diagnoses of these regions.
{"title":"Automated Breast Cancer Diagnosis Using Deep Learning and Region of Interest Detection (BC-DROID)","authors":"Richard Platania, Shayan Shams, Seungwon Yang, Jian Zhang, Kisung Lee, Seung-Jong Park","doi":"10.1145/3107411.3107484","DOIUrl":"https://doi.org/10.1145/3107411.3107484","url":null,"abstract":"Detection of suspicious regions in mammogram images and the subsequent diagnosis of these regions remains a challenging problem in the medical world. There still exists an alarming rate of misdiagnosis of breast cancer. This results in both over treatment through incorrect positive diagnosis of cancer and under treatment through overlooked cancerous masses. Convolutional neural networks have shown strong applicability to various image datasets, enabling detailed features to be learned from the data and, as a result, the ability to classify these images at extremely low error rates. In order to overcome the difficulty in diagnosing breast cancer from mammogram images, we propose our framework for automated breast cancer detection and diagnosis, called BC-DROID, which provides automated region of interest detection and diagnosis using convolutional neural networks. BC-DROID first pretrains based on physician-defined regions of interest in mammogram images. It then trains based on the full mammogram image. The resulting network is able to detect and classify regions of interest as cancerous or benign in one step. We demonstrate the accuracy of our framework's ability to both locate the regions of interest as well as diagnose them. Our framework achieves a detection accuracy of up to 90% and a classification accuracy of 93.5% (AUC of 92.315%). To the best of our knowledge, this is the first work enabling both automated detection and diagnosis of these areas in one step from full mammogram images. Using our framework's website, a user can upload a single mammogram image, visualize suspicious regions, and receive the automated diagnoses of these regions.","PeriodicalId":246388,"journal":{"name":"Proceedings of the 8th ACM International Conference on Bioinformatics, Computational Biology,and Health Informatics","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128880527","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The advent of high throughput technologies contributes to the rapid growth of molecular-level knowledge about human disease. However, existing disease taxonomies tend to focus on either physiological characterizations of disease or the organizational and billing needs of hospitals. Most fail to fully incorporate our rapidly increasing knowledge about molecular causes of disease. More modern disease taxonomies would presumably be built based on the combination of clinical, physiological, and molecular data. In this study, we analyzed our ability to infer disease relationships from molecular data alone. This approach may provide insights into how to ultimately build more modern taxonomies of disease
{"title":"Building a Molecular Taxonomy of Disease","authors":"Jisoo Park, Benjamin J. Hescott, D. Slonim","doi":"10.1145/3107411.3108236","DOIUrl":"https://doi.org/10.1145/3107411.3108236","url":null,"abstract":"The advent of high throughput technologies contributes to the rapid growth of molecular-level knowledge about human disease. However, existing disease taxonomies tend to focus on either physiological characterizations of disease or the organizational and billing needs of hospitals. Most fail to fully incorporate our rapidly increasing knowledge about molecular causes of disease. More modern disease taxonomies would presumably be built based on the combination of clinical, physiological, and molecular data. In this study, we analyzed our ability to infer disease relationships from molecular data alone. This approach may provide insights into how to ultimately build more modern taxonomies of disease","PeriodicalId":246388,"journal":{"name":"Proceedings of the 8th ACM International Conference on Bioinformatics, Computational Biology,and Health Informatics","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131302577","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Fabiana Lorenzi, Rodrigo G. da Rosa, A. Peres, G. Dorneles, André Peres, F. Ricci
Obesity is a complex disease that involves genetic factors, inflammatory patterns, resilience and psycho-social factors. An effective system which is able to recommend adequate training for obese subjects that starts a new protocol would enhance the quality and success of the rehabilitation of these subjects. This paper presents a case-based reasoning system that suggests the most effective type of physical training exercise for obese individuals. The presented system was validated by domain experts and the results of this analysis show that case-based reasoning is a viable approach that can help to improve life of obese people.
{"title":"Tailoring Training for Obese Individuals with Case-Based Reasoning","authors":"Fabiana Lorenzi, Rodrigo G. da Rosa, A. Peres, G. Dorneles, André Peres, F. Ricci","doi":"10.1145/3107411.3107467","DOIUrl":"https://doi.org/10.1145/3107411.3107467","url":null,"abstract":"Obesity is a complex disease that involves genetic factors, inflammatory patterns, resilience and psycho-social factors. An effective system which is able to recommend adequate training for obese subjects that starts a new protocol would enhance the quality and success of the rehabilitation of these subjects. This paper presents a case-based reasoning system that suggests the most effective type of physical training exercise for obese individuals. The presented system was validated by domain experts and the results of this analysis show that case-based reasoning is a viable approach that can help to improve life of obese people.","PeriodicalId":246388,"journal":{"name":"Proceedings of the 8th ACM International Conference on Bioinformatics, Computational Biology,and Health Informatics","volume":"118 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133746721","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Session details: Session 7: Advancing Algorithms and Methods I","authors":"Mukul S. Bansal","doi":"10.1145/3254550","DOIUrl":"https://doi.org/10.1145/3254550","url":null,"abstract":"","PeriodicalId":246388,"journal":{"name":"Proceedings of the 8th ACM International Conference on Bioinformatics, Computational Biology,and Health Informatics","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115354155","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
R. Farhoodi, Max Shelbourne, Rebecca Hsieh, Nurit Haspel, Brian Hutchinson, F. Jagodzinski
Predicting how a point mutation alters a protein's stability can guide drug design initiatives which aim to counter the effects of serious diseases. Mutagenesis studies give insights about the effects of amino acid substitutions, but such wet-lab work is prohibitive due to the time and costs needed to assess the consequences of even a single mutation. Computational methods for predicting the effects of a mutation are available, with promising accuracy rates. In this work we study the utility of several machine learning methods and their ability to predict the effects of mutations. We in silico generate mutant protein structures, and compute several rigidity metrics for each of them. Our approach does not require costly calculations of energy functions that rely on atomic-level statistical mechanics and molecular energetics. Our metrics are features for support vector regression, random forest, and deep neural network methods. We validate the effects of our in silico mutations against experimental Delta Delta G stability data. We attain Pearson Correlations upwards of 0.69.
{"title":"Predicting the Effect of Point Mutations on Protein Structural Stability","authors":"R. Farhoodi, Max Shelbourne, Rebecca Hsieh, Nurit Haspel, Brian Hutchinson, F. Jagodzinski","doi":"10.1145/3107411.3107492","DOIUrl":"https://doi.org/10.1145/3107411.3107492","url":null,"abstract":"Predicting how a point mutation alters a protein's stability can guide drug design initiatives which aim to counter the effects of serious diseases. Mutagenesis studies give insights about the effects of amino acid substitutions, but such wet-lab work is prohibitive due to the time and costs needed to assess the consequences of even a single mutation. Computational methods for predicting the effects of a mutation are available, with promising accuracy rates. In this work we study the utility of several machine learning methods and their ability to predict the effects of mutations. We in silico generate mutant protein structures, and compute several rigidity metrics for each of them. Our approach does not require costly calculations of energy functions that rely on atomic-level statistical mechanics and molecular energetics. Our metrics are features for support vector regression, random forest, and deep neural network methods. We validate the effects of our in silico mutations against experimental Delta Delta G stability data. We attain Pearson Correlations upwards of 0.69.","PeriodicalId":246388,"journal":{"name":"Proceedings of the 8th ACM International Conference on Bioinformatics, Computational Biology,and Health Informatics","volume":"43 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114498493","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：