Multi-omics integration for neuroblastoma clinical endpoint prediction

Francescatto, Margherita; Chierici, Marco; Dezfooli, Setareh Rezvan; Zandonà, Alessandro; Jurman, Giuseppe; Furlanello, Cesare

doi:10.1186/s13062-018-0207-8

Cited by 40 publications

(20 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Regarding the network architecture, models relying on four layer networks perform the best for both clinical outcomes (Table 3). This is in agreement with previous studies that have reported that such relatively small networks (i.e., with three or four layers) can efficiently predict clinical outcomes of kidney cancer patients [17] or can capture relevant features for survival analyses of a neuroblastoma cohort [12].…”

Section: Discussionsupporting

confidence: 92%

A deep neural network approach to predicting clinical outcomes of neuroblastoma patients

Tranchevent

Azuaje

Rajapakse

2019

Preprint

View full text Add to dashboard Cite

The availability of high-throughput omics datasets from large patient cohorts has allowed the development of methods that aim at predicting patient clinical outcomes, such as survival and disease recurrence. Such methods are also important to better understand the biological mechanisms underlying disease etiology and development, as well as treatment responses. Recently, different predictive models, relying on distinct algorithms (including Support Vector Machines and Random Forests) have been investigated. In this context, deep learning strategies are of special interest due to their demonstrated superior performance over a wide range of problems and datasets. One of the main challenges of such strategies is the "small n large p" problem. Indeed, omics datasets typically consist of small numbers of samples and large numbers of features relative to typical deep learning datasets. Neural networks usually tackle this problem through feature selection or by including additional constraints during the learning process.We propose to tackle this problem with a novel strategy that relies on a graph-based method for feature extraction, coupled with a deep neural network for clinical outcome prediction. The omics data are first represented as graphs whose nodes represent patients, and edges represent correlations between the patients' omics profiles. Topological features, such as centralities, are then extracted from these graphs for every node. Lastly, these features are used as input to train and test various classifiers.

show abstract

Section: Discussionsupporting

confidence: 92%

A deep neural network approach to predicting clinical outcomes of neuroblastoma patients

Tranchevent

Azuaje

Rajapakse

2019

Preprint

View full text Add to dashboard Cite

show abstract

“…All shallow and DL models (including class balancing experiments) were trained within the DAP previously developed by FBK within the MAQC-II and SEQC challenges [31,32], the U.S. FDA initiatives for reproducibility of biomarkers. Briefly, our DAP uses a 10 × 5−fold stratified CV on TR to get a ranked feature list and a set of classification metrics [33], including the MCC. Data were rescaled in the interval [ −1, 1] (for shallow learning) or centered and scaled to unit variance (for DL) before undergoing classification: rescaling parameters from TR were used for rescaling both TR and TS subsets, so to avoid information leakage.…”

Section: Predictive Modeling Strategymentioning

confidence: 99%

Predictability of drug-induced liver injury by machine learning

et al. 2020

Self Cite

View full text Add to dashboard Cite

Background: Drug-induced liver injury (DILI) is a major concern in drug development, as hepatotoxicity may not be apparent at early stages but can lead to life threatening consequences. The ability to predict DILI from in vitro data would be a crucial advantage. In 2018, the Critical Assessment Massive Data Analysis group proposed the CMap Drug Safety challenge focusing on DILI prediction. Methods and results: The challenge data included Affymetrix GeneChip expression profiles for the two cancer cell lines MCF7 and PC3 treated with 276 drug compounds and empty vehicles. Binary DILI labeling and a recommended train/test split for the development of predictive classification approaches were also provided. We devised three deep learning architectures for DILI prediction on the challenge data and compared them to random forest and multi-layer perceptron classifiers. On a subset of the data and for some of the models we additionally tested several strategies for balancing the two DILI classes and to identify alternative informative train/test splits. All the models were trained with the MAQC data analysis protocol (DAP), i.e., 10x5 cross-validation over the training set. In all the experiments, the classification performance in both cross-validation and external validation gave Matthews correlation coefficient (MCC) values below 0.2. We observed minimal differences between the two cell lines. Notably, deep learning approaches did not give an advantage on the classification performance. Discussion: We extensively tested multiple machine learning approaches for the DILI classification task obtaining poor to mediocre performance. The results suggest that the CMap expression data on the two cell lines MCF7 and PC3 are not sufficient for accurate DILI label prediction. Reviewers: This article was reviewed by Maciej Kandula and Paweł P. Labaj.

show abstract

“…Obviously, a great debate about this procedure and all early and late risks linked to it arose: blow out of the blind end, reactivation of CD in the excluded segment with abdominal pain and infections, deprivation of a large portion of the colon for water absorption. Eventually, bypass procedure was abandoned due to findings of adenocarcinoma occurring in the excluded segment [ 19 , 38 – 40 ]. Surgical resection of the diseased bowel emerged as the procedure of choice for most patients with CD of the terminal ileum or with ileo-colitis, including complicated cases [ 41 ].…”

Section: Surgical Management and CD Recurrencementioning

confidence: 99%

Pathophysiology of Crohn’s disease inflammation and recurrence

et al. 2020

View full text Add to dashboard Cite

Chron’s Disease is a chronic inflammatory intestinal disease, first described at the beginning of the last century. The disease is characterized by the alternation of periods of flares and remissions influenced by a complex pathogenesis in which inflammation plays a key role. Crohn’s disease evolution is mediated by a complex alteration of the inflammatory response which is characterized by alterations of the innate immunity of the intestinal mucosa barrier together with a remodeling of the extracellular matrix through the expression of metalloproteins and increased adhesion molecules expression, such as MAcCAM-1. This reshaped microenvironment enhances leucocytes migration in the sites of inflammation, promoting a TH1 response, through the production of cytokines such as IL-12 and TNF-α. IL-12 itself and IL-23 have been targeted for the medical treatment of CD. Giving the limited success of medical therapies, the treatment of the disease is invariably surgical. This review will highlight the role of inflammation in CD and describe the surgical approaches for the prevention of the almost inevitable recurrence.

show abstract

Multi-omics integration for neuroblastoma clinical endpoint prediction

Cited by 40 publications

References 22 publications

A deep neural network approach to predicting clinical outcomes of neuroblastoma patients

A deep neural network approach to predicting clinical outcomes of neuroblastoma patients

Predictability of drug-induced liver injury by machine learning

Pathophysiology of Crohn’s disease inflammation and recurrence

Contact Info

Product

Resources

About