D. Petrini, C. Shimizu, G. Valente, Guilherme Folgueira, Guilherme Apolinario Silva Novaes, M. H. Katayama, P. Serio, R. A. Roela, T. Tucunduva, M. A. K. Folgueira, Hae Yong Kim
{"title":"Abstract 181: High-accuracy breast cancer detection in mammography using EfficientNet and end-to-end training","authors":"D. Petrini, C. Shimizu, G. Valente, Guilherme Folgueira, Guilherme Apolinario Silva Novaes, M. H. Katayama, P. Serio, R. A. Roela, T. Tucunduva, M. A. K. Folgueira, Hae Yong Kim","doi":"10.1158/1538-7445.AM2021-181","DOIUrl":"https://doi.org/10.1158/1538-7445.AM2021-181","url":null,"abstract":"Background:Breast cancer (BC) is the second most common cancer among women. BC screening is usually based on mammography interpreted by radiologists. Recently, some researchers have used deep learning to automatically diagnose BC in mammography and so assist radiologists. The progress of BC detection algorithms can be measured by their performance on public datasets. The CBIS-DDSM is a widely used public dataset composed of scanned mammographies, equally divided into malignant and non-malignant (benign) images. Each image is accompanied by the segmentation of the lesion. Shen et al. (Nature Sci. Rep., 2019) presented a BC detection algorithm using an “end-to-end” approach to train deep neural networks. In this algorithm, a patch classifier is first trained to classify local image patches. The patch classifier9s weights are then used to initialize the whole image classifier, that is refined using datasets with the cancer status of the whole image. They achieved an AUC of 0.87 [0.84, 0.90] in classifying CBIS-DDSM images, using their best single-model, single-view breast classifier. They used ResNet (He et al., CVPR 2016) as the basis of their algorithm. Our hypothesis was that replacing the old ResNet with the modern EfficientNet (Tan et al., arXiv 2019) and MobileNetV2 (Sandler et al.,CVPR 2018) would result in greater accuracy. Methods:We tested many different models, to conclude that the best model is obtained using EfficientNet-B4 as the base model, with a MobileNetV2 block at the top, followed by a dense layer with two output categories. We trained the patch classifier using 52,528 patches with 224x224 pixels extracted from CBIS-DDSM. From each image, we extracted 20 patches: 10 patches containing the lesion and 10 from the background (without lesion). The patch classifier weights were then used to initialize the whole image classifier, that was trained using the end-to-end approach with CBIS-DDSM images resized to 1152x896 pixels, with data augmentation. The training was performed using a step learning rate of 1e-4 for the first 20 epochs then 1e-5 for the remaining 10 and batch size of 4, using 10-fold cross-validation. We used 81% of the dataset for training, 9% for validation and 10% for testing. Results:We obtained an AUC of 0.8963±0.06, using a single-model, single-view classifier and without test-time data augmentation. Conclusions:Using EfficientNet and MobileNetV2 as the basis of the BC detection algorithm (instead of ResNet), we obtained an improvement in classifying CBIS-DDSM images into malignant/non-malignant: AUC has increased from 0.87 to 0.896. Our AUC is also larger than other recent papers in the literature, such as Shu et al. (IEEE Trans Med. Image, 2020) that achieved an AUC of 0.838 in the same CBIS-DDSM dataset. Citation Format: Daniel G. Petrini, Carlos Shimizu, Gabriel V. Valente, Guilherme Folgueira, Guilherme A. Novaes, Maria L. Katayama, Pedro Serio, Rosimeire A. Roela, Tatiana C. Tucunduva, Maria Aparecida A. Folgu","PeriodicalId":73617,"journal":{"name":"Journal of bioinformatics and systems biology : Open access","volume":"85 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2021-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86986647","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}
Christie L. Husted, F. Aguet, C. Shea, A. Gower, William J. Mischler, Y. Koga, R. Hong, S. Dubinett, A. Spira, S. Mazzilli, E. Cerami, I. Leshchiner, M. Lenburg, G. Getz, J. Beane, Joshua D. Campbell
{"title":"Abstract 171: Cloud-based bulk and single-cell RNAseq pipelines in the Terra platform for the Lung PCA","authors":"Christie L. Husted, F. Aguet, C. Shea, A. Gower, William J. Mischler, Y. Koga, R. Hong, S. Dubinett, A. Spira, S. Mazzilli, E. Cerami, I. Leshchiner, M. Lenburg, G. Getz, J. Beane, Joshua D. Campbell","doi":"10.1158/1538-7445.AM2021-171","DOIUrl":"https://doi.org/10.1158/1538-7445.AM2021-171","url":null,"abstract":"","PeriodicalId":73617,"journal":{"name":"Journal of bioinformatics and systems biology : Open access","volume":"5 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2021-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86554885","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. Verma, Wei Wu, Neeraj Kumar, Elizabeth A. Yu, Won-Tak Choi, S. Umetsu, T. Bivona
{"title":"Abstract 3: Deep learning-based integration of esophageal cancer morphology with genomics","authors":"R. Verma, Wei Wu, Neeraj Kumar, Elizabeth A. Yu, Won-Tak Choi, S. Umetsu, T. Bivona","doi":"10.1158/1538-7445.AM2021-3","DOIUrl":"https://doi.org/10.1158/1538-7445.AM2021-3","url":null,"abstract":"","PeriodicalId":73617,"journal":{"name":"Journal of bioinformatics and systems biology : Open access","volume":"153 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2021-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86134801","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}
Chengyue Wu, D. Hormuth, F. Pineda, G. Karczmar, T. Yankeelov
{"title":"Abstract 222: Towards patient-specific optimization of neoadjuvant treatment protocols for breast cancer based on image-based fluid dynamics","authors":"Chengyue Wu, D. Hormuth, F. Pineda, G. Karczmar, T. Yankeelov","doi":"10.1158/1538-7445.AM2021-222","DOIUrl":"https://doi.org/10.1158/1538-7445.AM2021-222","url":null,"abstract":"","PeriodicalId":73617,"journal":{"name":"Journal of bioinformatics and systems biology : Open access","volume":"1 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2021-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74885438","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":"Abstract 168: Computational model for prediction of actionable drug combinations in cancer","authors":"Sairahul R Pentaparthi, Brandon Burgman, Song Yi","doi":"10.1158/1538-7445.AM2021-168","DOIUrl":"https://doi.org/10.1158/1538-7445.AM2021-168","url":null,"abstract":"","PeriodicalId":73617,"journal":{"name":"Journal of bioinformatics and systems biology : Open access","volume":"97 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2021-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78093613","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":"Abstract 196: Redefining cancer subtypes using multi-omics and deep learning","authors":"A. Akalin, B. Uyar, J. Ronen, V. Franke","doi":"10.1158/1538-7445.AM2021-196","DOIUrl":"https://doi.org/10.1158/1538-7445.AM2021-196","url":null,"abstract":"Cancer is a heterogeneous collection of diseases traditionally classified by the tissue of origin. The diversity of the molecular profiles of cancers has a big impact on the way patients are diagnosed and treated, how they respond to their prescribed treatments, the duration of survival after diagnosis, and factors such as remission, recurrence, or spread (metastasis) of the disease. While such diagnostic and prognostic outcomes are potentially predictable by taking a closer look into the changes of the genome, epigenome, transcriptome, proteome, and various other omics platforms, the contemporary cancer treatments still predominantly don9t make the best use of such multi-omics profiling of patient samples. Therefore, multi-omics profiling of cancers holds great potential to define a molecularly coherent subtype definition of cancers in order to achieve the eventual goal of matching the best possible treatment to the subgroup of patients. However, the current subtypes from consortiums such as TCGA have been defined by heterogeneous methods and molecular markers by different teams. A subset of these studies have not attempted to characterize molecular subtypes, but rather taken histopathologically defined subtypes as the gold standard and tried to characterize molecular features of these subtypes. Here we evaluate TCGA cancer subtypes based on the molecular profile coherence score. This novel metric combines survival statistics, pathways information, tumor purity estimates, and mutational signatures. We expect that subtypes that are patient subgroups should display molecular signature homogeneity. We evaluate TCGA subtypes from 21 cancers using these criteria and compare the subtypes with our own definition using multi-omics data in a deep learning framework. We have refined the several subtypes from multiple cancers towards more molecularly coherent patient subgroups. Citation Format: Altuna Akalin, Bora Uyar, Jonathan Ronen, Vedran Franke. Redefining cancer subtypes using multi-omics and deep learning [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 196.","PeriodicalId":73617,"journal":{"name":"Journal of bioinformatics and systems biology : Open access","volume":"284 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2021-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"76851005","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":"Abstract 202: Identification of survival associated hub genes in prostate cancer patients from the TCGA database","authors":"N. Reyes, R. Tiwari, J. Geliebter","doi":"10.1158/1538-7445.AM2021-202","DOIUrl":"https://doi.org/10.1158/1538-7445.AM2021-202","url":null,"abstract":"Background: Prostate cancer is the most frequently diagnosed malignancy and the fourth leading cause of cancer-related death in the global male population. Although the disease has a relatively low mortality rate with some patients surviving for 10-20 years after treatment, others respond poorly to treatment and die of metastatic disease within 2-3 years. Therefore, there is an urgent need to develop strategies to identify patients with clinically significant prostate cancer requiring aggressive treatment to improve survival, while sparing others unnecessary side effects. The purpose of this study was to identify survival associated genes in prostate cancer patients from the TCGA database using bioinformatics tools. Methods: Data from prostate cancer patients in the TCGA database were divided into two study groups: a high and a low expression group, relative to the median expression. The Gene Expression Profiling Interactive Analysis (GEPIA2) tool was used for the identification of the most differential survival genes. Metascape bioinformatics tool was subsequently used for clustering of genes based on processes, pathway enrichment analysis, and construction of Protein-Protein Interaction (PPI) network. Metascape was also used for molecular Complex Detection (MCODE) to identify the genes with the highest degree of connection, known as hub genes, and to screen modules of the PPI network. Results: Bioinformatics analysis allowed the identification of 361 genes whose expression levels were significantly associated with overall survival in prostate cancer patients from the TCGA. Survival associated genes were primarily enriched in mRNA processing, DNA repair, ncRNA processing, DNA replication, macromolecule methylation, among others. The 12 most connected genes were selected as hub genes and Kaplan-Meier analysis was used to verify survival associated with this set of genes. Hub genes included several splicing factors and components of the processing machinery of cellular pre-mRNAs. Conclusions: These hub genes may reveal basic mechanisms underlying the development of clinically relevant prostate cancer and contribute to the identification of novel markers for prognosis of this cancer. Citation Format: Niradiz Reyes, Raj Tiwari, Jan Geliebter. Identification of survival associated hub genes in prostate cancer patients from the TCGA database [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 202.","PeriodicalId":73617,"journal":{"name":"Journal of bioinformatics and systems biology : Open access","volume":"201 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2021-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"76923369","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}
Zinaida Perova, Csaba Halmagyi, Abayomi Mosaku, N. Conte, J. Mason, Alex Follette, Ross Thorne, Mauricio Martinez, S. Neuhauser, D. Begley, D. Krupke, H. Parkinson, T. Meehan, C. Bult
{"title":"Abstract LB017: PDX Finder: An open and global catalogue of patient-derived xenograft models","authors":"Zinaida Perova, Csaba Halmagyi, Abayomi Mosaku, N. Conte, J. Mason, Alex Follette, Ross Thorne, Mauricio Martinez, S. Neuhauser, D. Begley, D. Krupke, H. Parkinson, T. Meehan, C. Bult","doi":"10.1158/1538-7445.AM2021-LB017","DOIUrl":"https://doi.org/10.1158/1538-7445.AM2021-LB017","url":null,"abstract":"Patient-derived tumor xenograft (PDX) models are a critical oncology platform for cancer research, drug development and personalized medicine. Because of the heterogeneous nature of PDXs repositories, finding models of interest is a challenge. The Jackson Laboratory and EMBL-EBI are developing PDX Finder, the world9s largest open PDX database containing millions of phenomic information from over 4300 models (www.pdxfinder.org, PMID: 30535239). In support of this initiative, we developed the PDX Minimal Information standard (PDX-MI) which defines metadata necessary to describe models (PMID: 29092942). Within PDX Finder, critical attributes like diagnosis, drug names or genes are harmonized into a cohesive ontological data model based on PDX-MI. An intuitive search and faceted search interface allow users to select models based on clinical/PDX attributes, tumor markers, dataset availability and/or drug dosing results. We provide PDX, patient, drug and molecular data detail pages where all available information can be browsed and downloaded. To further facilitate user9s model selection, we are linking key external resources like publication platforms and cancer-specific annotation tools enabling exploration and prioritization of PDX variation data (COSMIC, CIViC, OncoMx, OpenCRAVAT). Links to originating resource protocols and contact information are provided, facilitating data understanding and further collaboration. Alongside database development activities, PDX Finder has undertaken activities to tackle areas of standards and tool development, data integration and outreach. PDX Finder provides key expertise and software components to support several worldwide consortia including PDXNet, PDMR and EurOPDX. We are driving the development of, and promoting the use of descriptive standards to facilitate data interoperability and promote global sharing of models. Our standard has become established in the community for data exchange, adopted by PDX providers, consortia, and informatics tools integrating PDX data. It has been re-used by different initiatives in the context of data collection and data modeling allowing adherence to the FAIR data principles - Findability, Accessibility, Interoperability and Reusability. PDX Finder is increasing awareness of PDX models, facilitating data integration, and enabling international collaboration, maximizing the investment in, and translational capabilities of these important models of human cancer. PDX Finder is freely available under an Apache 2 license (github.com/pdxfinder). Work supported by NCI U24 CA204781 01 (ended 31Aug2020), U24 CA253539, and R01 CA089713. Citation Format: Zinaida Perova, Csaba Halmagyi, Abayomi Mosaku, Nathalie Conte, Jeremy Mason, Alex Follette, Ross Thorne, Mauricio Martinez, Steven Neuhauser, Dale Begley, Debra Krupke, Helen Parkinson, Terrence Meehan, Carol Bult. PDX Finder: An open and global catalogue of patient-derived xenograft models [abstract]. In: Proceedings of the America","PeriodicalId":73617,"journal":{"name":"Journal of bioinformatics and systems biology : Open access","volume":"26 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2021-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"73803204","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":"Abstract 154: Improving MultiOmyxTMAnalytics cell classification workflow efficiency by Invariant Information Clustering on historical data","authors":"V. Reddy, Nicholas Stavrou, M. Nagy, Q. Au","doi":"10.1158/1538-7445.AM2021-154","DOIUrl":"https://doi.org/10.1158/1538-7445.AM2021-154","url":null,"abstract":"","PeriodicalId":73617,"journal":{"name":"Journal of bioinformatics and systems biology : Open access","volume":"25 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2021-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"73476745","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}
Manuel García-Quismondo, O. Elemento, Neel S. Madhukar, Coryandar Gilvary
{"title":"Abstract 219: Identifying genetic interactions resulting form diverse biological mechanisms to inform cancer drug development","authors":"Manuel García-Quismondo, O. Elemento, Neel S. Madhukar, Coryandar Gilvary","doi":"10.1158/1538-7445.AM2021-219","DOIUrl":"https://doi.org/10.1158/1538-7445.AM2021-219","url":null,"abstract":"","PeriodicalId":73617,"journal":{"name":"Journal of bioinformatics and systems biology : Open access","volume":"17 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2021-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86640461","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}