Tamsen Dunn, G. Berry, Dorothea Emig-Agius, Yu Jiang, A. Iyer, N. Udar, Michael P. Strömberg
A method for robustly and accurately detecting rare DNA mutations in tumor samples is critical to cancer research. Because many clinical tissue repositories have only FFPE-degraded tumor samples, and no matched normal sample from healthy tissue available, being able to discriminate low frequency mutations from background noise in the absence of a matched normal sample is of particular importance to research. Current state of the art variant callers such as GATK and VarScan focus on germline variant calling (used for detecting inherited mutations following a Mendelian inheritance pattern) or, in the case of FreeBayes and MuTect, focus on tumor-normal joint variant calling (using the normal sample to help discriminate low frequency somatic mutations from back ground noise). We present Pisces, a tumor-only variant caller exclusively developed at Illumina for detecting low frequency mutations from next generation sequencing data. Pisces has been an integral part of the Illumina Truseq Amplicon workflow since 2012, and is available on BaseSpace and on the MiSeq sequencing platforms. Pisces has been available to the public on github, since 2015. (https://github.com/Illumina/Pisces) Since that time, the Pisces variant calling team have continued to develop Pisces, and have made available a suite of variant calling tools, including a ReadStitcher, Variant Phaser, and Variant Quality Recalibration tool, to be used along with the core variant caller, Pisces. Here, we describe the Pisces variant calling tools and core algorithms. We describe the common use cases for Pisces (not necessarily restricted to somatic variant calling). We also evaluate Pisces performance on somatic and germline datasets, both from the titration of well characterized samples, and from a corpus of 500 FFPE-treated clinical trial tumor samples, against other variant callers. Our results show that Pisces gives highly accurate results in a variety of contexts. We recommend Pisces for amplicon somatic and germline variant calling.
{"title":"Pisces: An Accurate and Versatile Single Sample Somatic and Germline Variant Caller","authors":"Tamsen Dunn, G. Berry, Dorothea Emig-Agius, Yu Jiang, A. Iyer, N. Udar, Michael P. Strömberg","doi":"10.1145/3107411.3108203","DOIUrl":"https://doi.org/10.1145/3107411.3108203","url":null,"abstract":"A method for robustly and accurately detecting rare DNA mutations in tumor samples is critical to cancer research. Because many clinical tissue repositories have only FFPE-degraded tumor samples, and no matched normal sample from healthy tissue available, being able to discriminate low frequency mutations from background noise in the absence of a matched normal sample is of particular importance to research. Current state of the art variant callers such as GATK and VarScan focus on germline variant calling (used for detecting inherited mutations following a Mendelian inheritance pattern) or, in the case of FreeBayes and MuTect, focus on tumor-normal joint variant calling (using the normal sample to help discriminate low frequency somatic mutations from back ground noise). We present Pisces, a tumor-only variant caller exclusively developed at Illumina for detecting low frequency mutations from next generation sequencing data. Pisces has been an integral part of the Illumina Truseq Amplicon workflow since 2012, and is available on BaseSpace and on the MiSeq sequencing platforms. Pisces has been available to the public on github, since 2015. (https://github.com/Illumina/Pisces) Since that time, the Pisces variant calling team have continued to develop Pisces, and have made available a suite of variant calling tools, including a ReadStitcher, Variant Phaser, and Variant Quality Recalibration tool, to be used along with the core variant caller, Pisces. Here, we describe the Pisces variant calling tools and core algorithms. We describe the common use cases for Pisces (not necessarily restricted to somatic variant calling). We also evaluate Pisces performance on somatic and germline datasets, both from the titration of well characterized samples, and from a corpus of 500 FFPE-treated clinical trial tumor samples, against other variant callers. Our results show that Pisces gives highly accurate results in a variety of contexts. We recommend Pisces for amplicon somatic and germline variant calling.","PeriodicalId":246388,"journal":{"name":"Proceedings of the 8th ACM International Conference on Bioinformatics, Computational Biology,and Health Informatics","volume":"8 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":"130083217","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}
Bruno Dantas, Calmenelias Fleitas, Alexandre Almeida, J. Forja, Alexandre P. Francisco, José Simão, Cátia Vaz
Biosciences have been revolutionised by NGS technologies in last years, leading to new perspectives in medical, industrial and environmental applications. And although our motivation comes from biosciences, the following is true for many areas of science: published results are usually hard to reproduce, delaying the adoption of new methodologies and hindering innovation. Even if data and tools are freely available, pipelines for data analysis are in general barely described and their setup is far from trivial. NGSPipes addresses these issues reducing the efforts necessary to define, build and deploy pipelines, either at a local workstation or in the cloud. NGSPipes framework is freely available at http://ngspipes.github.io/.
{"title":"NGSPipes: Fostering Reproducibility and Scalability in Biosciences","authors":"Bruno Dantas, Calmenelias Fleitas, Alexandre Almeida, J. Forja, Alexandre P. Francisco, José Simão, Cátia Vaz","doi":"10.1145/3107411.3108213","DOIUrl":"https://doi.org/10.1145/3107411.3108213","url":null,"abstract":"Biosciences have been revolutionised by NGS technologies in last years, leading to new perspectives in medical, industrial and environmental applications. And although our motivation comes from biosciences, the following is true for many areas of science: published results are usually hard to reproduce, delaying the adoption of new methodologies and hindering innovation. Even if data and tools are freely available, pipelines for data analysis are in general barely described and their setup is far from trivial. NGSPipes addresses these issues reducing the efforts necessary to define, build and deploy pipelines, either at a local workstation or in the cloud. NGSPipes framework is freely available at http://ngspipes.github.io/.","PeriodicalId":246388,"journal":{"name":"Proceedings of the 8th ACM International Conference on Bioinformatics, Computational Biology,and Health Informatics","volume":"228 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":"114592650","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}
Reconstructing 3D structure of a genome from chromosomal conformation capturing data such as Hi-C data has emerged as an important problem in bioinformatics and computational biology in the recent years. In this talk, I will present our latest method that uses Lorentzian function to describe distance restraints between chromosomal regions, which will be used to guide the reconstruction of 3D structures of individual chromosomes and an entire genome. The method is more robust against noisy distance restraints derived from Hi-C data than traditional objective functions such as squared error function and Gaussian probabilistic function. The method can handle both intra- and inter-chromosomal contacts effectively to build 3D structures of a big genome such as the human genome consisting of a number of chromosomes, which are not possible with most existing methods. We have released the Java source code that implements the method (called LorDG) at GitHub (https://github.com/BDM-Lab/LorDG), which is being used by the community to model 3D genome structures. We are currently further improving the method to build very high-resolution (e.g. 1KB base pair) 3D genome and chromosome models.
{"title":"3D Genome Structure Modeling by Lorentzian Objective Function","authors":"Tuan Trieu, Jianlin Cheng","doi":"10.1145/3107411.3107455","DOIUrl":"https://doi.org/10.1145/3107411.3107455","url":null,"abstract":"Reconstructing 3D structure of a genome from chromosomal conformation capturing data such as Hi-C data has emerged as an important problem in bioinformatics and computational biology in the recent years. In this talk, I will present our latest method that uses Lorentzian function to describe distance restraints between chromosomal regions, which will be used to guide the reconstruction of 3D structures of individual chromosomes and an entire genome. The method is more robust against noisy distance restraints derived from Hi-C data than traditional objective functions such as squared error function and Gaussian probabilistic function. The method can handle both intra- and inter-chromosomal contacts effectively to build 3D structures of a big genome such as the human genome consisting of a number of chromosomes, which are not possible with most existing methods. We have released the Java source code that implements the method (called LorDG) at GitHub (https://github.com/BDM-Lab/LorDG), which is being used by the community to model 3D genome structures. We are currently further improving the method to build very high-resolution (e.g. 1KB base pair) 3D genome and chromosome models.","PeriodicalId":246388,"journal":{"name":"Proceedings of the 8th ACM International Conference on Bioinformatics, Computational Biology,and Health Informatics","volume":"73 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":"133573467","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 energy landscape underscores the inherent nature of proteins as dynamic systems interconverting between structures with varying energies. The protein energy landscape contains much of the information needed to characterize protein equilibrium dynamics and relate it to function. It is now possible to reconstruct energy landscapes of medium-size proteins with sufficient prior structure data. These developments turn the focus to tools for analysis and comparison of energy landscapes as a means of formulating hypotheses on the impact of sequence mutations on (dys)function via altered landscape features. We present such a method here and provide a detailed evaluation of its capabilities on an enzyme central to human biology. The work presented here opens up an interesting avenue into automated analysis and summarization of landscapes that yields itself to machine learning approaches at the energy landscape level.
{"title":"Statistical Analysis of Computed Energy Landscapes to Understand Dysfunction in Pathogenic Protein Variants","authors":"Wanli Qiao, T. Maximova, E. Plaku, Amarda Shehu","doi":"10.1145/3107411.3107499","DOIUrl":"https://doi.org/10.1145/3107411.3107499","url":null,"abstract":"The energy landscape underscores the inherent nature of proteins as dynamic systems interconverting between structures with varying energies. The protein energy landscape contains much of the information needed to characterize protein equilibrium dynamics and relate it to function. It is now possible to reconstruct energy landscapes of medium-size proteins with sufficient prior structure data. These developments turn the focus to tools for analysis and comparison of energy landscapes as a means of formulating hypotheses on the impact of sequence mutations on (dys)function via altered landscape features. We present such a method here and provide a detailed evaluation of its capabilities on an enzyme central to human biology. The work presented here opens up an interesting avenue into automated analysis and summarization of landscapes that yields itself to machine learning approaches at the energy landscape level.","PeriodicalId":246388,"journal":{"name":"Proceedings of the 8th ACM International Conference on Bioinformatics, Computational Biology,and Health Informatics","volume":"27 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":"116564511","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}
Shaojun Tang, Suthee Rapisuwon, A. Wellstein, Subha Madhavan
Successful treatment of cancers with Immune Checkpoint Inhibitors (ICIs) has been associated with the mutational load of tumors. The biological rationale for this association between mutational load and ICI response is that neoantigens are generated by mutations in protein coding sequences that provide a steady flow of neoantigens to prime the immune system for the production of antigen-specific tumor-infiltrating lymphocytes (TILs). It is thought that mutant protein fragments will lead to altered MHC/peptide recognition and immune cell activation; ICI treatment enhances TIL functionality. Neoantigens are also relevant for an alternative, cell-based immunotherapeutic approach, i.e. Adoptive Cell Transfer (ACT). This concept of neoantigens derived from DNA mutations has led to an intense line of investigation to uncover relevant neoantigens. However, there has been mixed success with the current neoantigen discovery approach based on DNA mutation analysis of tumor samples by exome sequencing of genomic DNA. The current concept of neoantigens derived from mutant DNA ignores an alternative mechanism that can also generate neoantigens in cancers: Posttranscriptional editing of primary RNA. Here we propose to use full-length Single Molecule Real Time (SMRT) RNAseq to uncover pathologically edited mRNAs in cancers and complement the discovery of pathologic mRNA. We will discuss the respective algorithms and propose the combination with identification of candidate neoantigen peptides by mass spectrometry.
{"title":"Tumor Neoantigens Derived from RNA Sequencing Analysis","authors":"Shaojun Tang, Suthee Rapisuwon, A. Wellstein, Subha Madhavan","doi":"10.1145/3107411.3108210","DOIUrl":"https://doi.org/10.1145/3107411.3108210","url":null,"abstract":"Successful treatment of cancers with Immune Checkpoint Inhibitors (ICIs) has been associated with the mutational load of tumors. The biological rationale for this association between mutational load and ICI response is that neoantigens are generated by mutations in protein coding sequences that provide a steady flow of neoantigens to prime the immune system for the production of antigen-specific tumor-infiltrating lymphocytes (TILs). It is thought that mutant protein fragments will lead to altered MHC/peptide recognition and immune cell activation; ICI treatment enhances TIL functionality. Neoantigens are also relevant for an alternative, cell-based immunotherapeutic approach, i.e. Adoptive Cell Transfer (ACT). This concept of neoantigens derived from DNA mutations has led to an intense line of investigation to uncover relevant neoantigens. However, there has been mixed success with the current neoantigen discovery approach based on DNA mutation analysis of tumor samples by exome sequencing of genomic DNA. The current concept of neoantigens derived from mutant DNA ignores an alternative mechanism that can also generate neoantigens in cancers: Posttranscriptional editing of primary RNA. Here we propose to use full-length Single Molecule Real Time (SMRT) RNAseq to uncover pathologically edited mRNAs in cancers and complement the discovery of pathologic mRNA. We will discuss the respective algorithms and propose the combination with identification of candidate neoantigen peptides by mass spectrometry.","PeriodicalId":246388,"journal":{"name":"Proceedings of the 8th ACM International Conference on Bioinformatics, Computational Biology,and Health Informatics","volume":"119 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":"123487449","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 rapid accumulation of macromolecular structures presents a unique set of challenges and opportunities in the analysis, comparison, modeling, and prediction of macromolecular structures and interactions. This workshop aims to bring together researchers with expertise in bioinformatics, computational biology, structural biology, data mining, optimization and high performance computing to discuss new results, techniques, and research problems in computational structural bioinformatics.
{"title":"The 2017 Computational Structural Bioinformatics Workshop: CSBW 2017","authors":"Nurit Haspel, Amarda Shehu, Kevin Molloy","doi":"10.1145/3107411.3108166","DOIUrl":"https://doi.org/10.1145/3107411.3108166","url":null,"abstract":"The rapid accumulation of macromolecular structures presents a unique set of challenges and opportunities in the analysis, comparison, modeling, and prediction of macromolecular structures and interactions. This workshop aims to bring together researchers with expertise in bioinformatics, computational biology, structural biology, data mining, optimization and high performance computing to discuss new results, techniques, and research problems in computational structural bioinformatics.","PeriodicalId":246388,"journal":{"name":"Proceedings of the 8th ACM International Conference on Bioinformatics, Computational Biology,and Health Informatics","volume":"201 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":"122349046","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}
Genome-wide ChIP-seq analysis of transcription factor binding and histone marks has uncovered large regulatory domains, known as super-enhancers, that consist of clusters of active enhancers within 12.5 kb of each other. Super-enhancers are characterized by high abundance of H3K27ac histone marks, disproportionate binding of master regulators and coactivators, and drive the expression of important cell identity genes. The algorithm, Rank Ordering of Super-Enhancers (ROSE), was developed to identify SEs based on their characteristics and has been extensively used, on various cell types. Less attention has been aimed at understanding how super-enhancers change in different cellular contexts, and in particular, during the differentiation of stem cells. We use ROSE, in conjunction with other tools, to investigate the dynamics of super-enhancers across myogenesis. Using ChIP-seq data for various transcription factors and stage-matched RNA-seq data, we characterize several super-enhancer regions and their associated genes in myoblasts and myotubes, finding them to be largely stage-specific.
转录因子结合和组蛋白标记的全基因组ChIP-seq分析发现了被称为超级增强子的大型调控结构域，这些结构域由12.5 kb以内的活性增强子簇组成。超级增强子的特点是高丰度的H3K27ac组蛋白标记，与主调控因子和共激活因子不成比例的结合，并驱动重要细胞身份基因的表达。超级增强子Rank Ordering of super - enhancer (ROSE)算法是一种基于se特征的识别算法，已被广泛应用于各种细胞类型。很少有人关注超级增强子如何在不同的细胞环境中发生变化，特别是在干细胞分化过程中。我们使用ROSE，结合其他工具，来研究超级增强剂在肌肉形成过程中的动态。利用各种转录因子的ChIP-seq数据和阶段匹配的RNA-seq数据，我们表征了成肌细胞和肌管中的几个超级增强子区域及其相关基因，发现它们在很大程度上是阶段特异性的。
{"title":"Super-enhancer Dynamics Throughout Myogenesis","authors":"Basma Abdelkarim, T. Perkins","doi":"10.1145/3107411.3108187","DOIUrl":"https://doi.org/10.1145/3107411.3108187","url":null,"abstract":"Genome-wide ChIP-seq analysis of transcription factor binding and histone marks has uncovered large regulatory domains, known as super-enhancers, that consist of clusters of active enhancers within 12.5 kb of each other. Super-enhancers are characterized by high abundance of H3K27ac histone marks, disproportionate binding of master regulators and coactivators, and drive the expression of important cell identity genes. The algorithm, Rank Ordering of Super-Enhancers (ROSE), was developed to identify SEs based on their characteristics and has been extensively used, on various cell types. Less attention has been aimed at understanding how super-enhancers change in different cellular contexts, and in particular, during the differentiation of stem cells. We use ROSE, in conjunction with other tools, to investigate the dynamics of super-enhancers across myogenesis. Using ChIP-seq data for various transcription factors and stage-matched RNA-seq data, we characterize several super-enhancer regions and their associated genes in myoblasts and myotubes, finding them to be largely stage-specific.","PeriodicalId":246388,"journal":{"name":"Proceedings of the 8th ACM International Conference on Bioinformatics, Computational Biology,and Health Informatics","volume":"14 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":"125897192","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 17: Biological Modeling","authors":"N. Yanamala","doi":"10.1145/3254560","DOIUrl":"https://doi.org/10.1145/3254560","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":"127468351","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}
Yanshan Wang, Sijia Liu, M. Rastegar-Mojarad, Liwei Wang, F. Shen, Fei Liu, Hongfang Liu
Drug-drug interaction (DDI) is an unexpected change in a drug's effect on the human body when the drug and a second drug are co-prescribed and taken together. As many DDIs are frequently reported in biomedical literature, it is important to mine DDI information from literature to keep DDI knowledge up to date. One of the SemEval challenges in the year 2011 and 2013 was designed to tackle the task where the best system achieved an F1 score of 0.80. In this paper, we propose to utilize dependency embeddings and Abstract Meaning Representation (AMR) embeddings as features for extracting DDIs. Our contribution is two-fold. First, we employed dependency embeddings, previously shown effective for sentence classification, for DDI extraction. The dependency embeddings incorporated structural syntactic contexts into the embeddings, which were not present in the conventional word embeddings. Second, we proposed a novel syntactic embedding approach using AMR. AMR aims to abstract away from syntactic idiosyncrasies and attempts to capture only the core meaning of a sentence, which could potentially improve DDI extraction from sentences. Two classifiers (Support Vector Machine and Random Forest) taking these embedding features as input were evaluated on the DDIExtraction 2013 challenge corpus. The experimental results show the effectiveness of dependency and AMR embeddings in the DDI extraction task. The best performance was obtained by combining word, dependency and AMR embeddings (F1 score=0.84).
{"title":"Dependency and AMR Embeddings for Drug-Drug Interaction Extraction from Biomedical Literature","authors":"Yanshan Wang, Sijia Liu, M. Rastegar-Mojarad, Liwei Wang, F. Shen, Fei Liu, Hongfang Liu","doi":"10.1145/3107411.3107426","DOIUrl":"https://doi.org/10.1145/3107411.3107426","url":null,"abstract":"Drug-drug interaction (DDI) is an unexpected change in a drug's effect on the human body when the drug and a second drug are co-prescribed and taken together. As many DDIs are frequently reported in biomedical literature, it is important to mine DDI information from literature to keep DDI knowledge up to date. One of the SemEval challenges in the year 2011 and 2013 was designed to tackle the task where the best system achieved an F1 score of 0.80. In this paper, we propose to utilize dependency embeddings and Abstract Meaning Representation (AMR) embeddings as features for extracting DDIs. Our contribution is two-fold. First, we employed dependency embeddings, previously shown effective for sentence classification, for DDI extraction. The dependency embeddings incorporated structural syntactic contexts into the embeddings, which were not present in the conventional word embeddings. Second, we proposed a novel syntactic embedding approach using AMR. AMR aims to abstract away from syntactic idiosyncrasies and attempts to capture only the core meaning of a sentence, which could potentially improve DDI extraction from sentences. Two classifiers (Support Vector Machine and Random Forest) taking these embedding features as input were evaluated on the DDIExtraction 2013 challenge corpus. The experimental results show the effectiveness of dependency and AMR embeddings in the DDI extraction task. The best performance was obtained by combining word, dependency and AMR embeddings (F1 score=0.84).","PeriodicalId":246388,"journal":{"name":"Proceedings of the 8th ACM International Conference on Bioinformatics, Computational Biology,and Health Informatics","volume":"1 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":"128810002","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}
Sangkyun Lee, S. Kerns, B. Rosenstein, H. Ostrer, J. Deasy, J. Oh
Urinary toxicity after radiotherapy (RT) limits the quality of life of prostate cancer patients, and clinically actionable prediction has yet to be achieved. We aim to exploit genome-wide variants to accurately identify patients at higher congenital toxicity risk. We applied preconditioned random forest regression (PRFR) to predict four urinary symptoms. For a weak stream endpoint, the PRFR model achieved an area under the curve (AUC) of 0.7 on holdout validation. Preconditioning enhanced the performance of random forest. Gene ontology (GO) analysis showed that neurogenic biological processes are associated with the toxicity. Upon further validation, the predictive model can be used to potentially benefit the health of prostate cancer patients treated with radiotherapy.
{"title":"Preconditioned Random Forest Regression: Application to Genome-Wide Study for Radiotherapy Toxicity Prediction","authors":"Sangkyun Lee, S. Kerns, B. Rosenstein, H. Ostrer, J. Deasy, J. Oh","doi":"10.1145/3107411.3108201","DOIUrl":"https://doi.org/10.1145/3107411.3108201","url":null,"abstract":"Urinary toxicity after radiotherapy (RT) limits the quality of life of prostate cancer patients, and clinically actionable prediction has yet to be achieved. We aim to exploit genome-wide variants to accurately identify patients at higher congenital toxicity risk. We applied preconditioned random forest regression (PRFR) to predict four urinary symptoms. For a weak stream endpoint, the PRFR model achieved an area under the curve (AUC) of 0.7 on holdout validation. Preconditioning enhanced the performance of random forest. Gene ontology (GO) analysis showed that neurogenic biological processes are associated with the toxicity. Upon further validation, the predictive model can be used to potentially benefit the health of prostate cancer patients treated with radiotherapy.","PeriodicalId":246388,"journal":{"name":"Proceedings of the 8th ACM International Conference on Bioinformatics, Computational Biology,and Health Informatics","volume":"3 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":"129263555","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：