Orchestrating and sharing large multimodal data for transparent and reproducible research

Mammoliti, Anthony; Смирнов, Петр; Nakano, Masahiro; Safikhani, Zhaleh; Eeles, Christopher; Seo, Hiromi; Nair, Sisira Kadambat; Mer, Arvind Singh; Ho, Chantal; Beri, Gangesh; Kusko, Rebecca; Haibe‐Kains, Benjamin

doi:10.1101/2020.09.18.303842

Cited by 6 publications

(6 citation statements)

References 69 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This includes adhering to the FAIR data principles, along with ensuring that there is a standardized manner in which the pipelines are executed and the data is hosted. To address this important issue, we used ORCESTRA (https://orcestra.ca/), a platform that allows researchers to process biomedical data into unified data objects in a reproducible and transparent manner, where data provenance is tracked (26). At the heart of ORCESTRA is Pachyderm (https://www.pachyderm.com/), an open-source data versioning tool used to execute pipelines processing the molecular and compound screening data, and packaging the datasets into R objects called PharmacoSets (PSets), which are implemented by the PharmacoGx package (27).…”

Section: Implementation Of Reproducible Pipelinesmentioning

confidence: 99%

PharmacoDB 2.0 : Improving scalability and transparency ofin vitropharmacogenomics analysis

Feizi

Nair

Смирнов

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Cancer pharmacogenomics studies provide valuable insights into disease progression and associations between genomic features and drug response. PharmacoDB integrates multiple cancer pharmacogenomics datasets profiling approved and investigational drugs across cell lines from diverse tissue types. The web-application enables users to efficiently navigate across datasets, view and compare drug dose-response data for a specific drug-cell line pair. In the new version of PharmacoDB (version 2.0, https://pharmacodb.ca/), we present: (i) new datasets such as NCI-60, the Profiling Relative Inhibition Simultaneously in Mixtures (PRISM) dataset, as well as updated data from the Genomics of Drug Sensitivity in Cancer (GDSC) and the Genentech Cell Line Screening Initiative (gCSI); (ii) implementation of FAIR data pipelines using ORCESTRA and PharmacoDI; (iii) enhancements to drug response analysis such as tissue distribution of dose-response metrics and biomarker analysis; (iv) improved connectivity to drug and cell line databases in the community. The web interface has been rewritten using a modern technology stack to ensure scalability and standardization to accommodate growing pharmacogenomics datasets. PharmacoDB 2.0 is a valuable tool for mining pharmacogenomics datasets, comparing and assessing drug response phenotypes of cancer models.

show abstract

Section: Implementation Of Reproducible Pipelinesmentioning

confidence: 99%

PharmacoDB 2.0 : Improving scalability and transparency ofin vitropharmacogenomics analysis

Feizi

Nair

Смирнов

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…We obtained all cell line datasets from the ORCESTRA platform (Mammoliti et al 2020) which stores pharmacogenomics datasets in PharmacoSet (PSet) R objects. Samples with missing values were removed from both the gene expression and drug response data.…”

Section: Data Preprocessingmentioning

confidence: 99%

Out-of-Distribution Generalization from Labeled and Unlabeled Gene Expression Data for Drug Response Prediction

Sharifi-Noghabi

Harjandi

Zolotareva

et al. 2021

Preprint

View full text Add to dashboard Cite

Data discrepancy between preclinical and clinical datasets poses a major challenge for accurate drug response prediction based on gene expression data. Different methods of transfer learning have been proposed to address this data discrepancy. These methods generally use cell lines as source domains and patients, patient-derived xenografts, or other cell lines as target domains. However, they assume that they have access to the target domain during training or fine-tuning and they can only take labeled source domains as input. The former is a strong assumption that is not satisfied during deployment of these models in the clinic. The latter means these methods rely on labeled source domains which are of limited size. To avoid this assumption, we formulate drug response prediction as an out-of-distribution generalization problem which does not assume that the target domain is accessible during training. Moreover, to exploit unlabeled source domain data, which tends to be much more plentiful than labeled data, we adopt a semi-supervised approach. We propose Velodrome, a semi-supervised method of out-of-distribution generalization that takes labeled and unlabeled data from different resources as input and makes generalizable predictions. Velodrome achieves this goal by introducing an objective function that combines a supervised loss for accurate prediction, an alignment loss for generalization, and a consistency loss to incorporate unlabeled samples. Our experimental results demonstrate that Velodrome outperforms state-of-the-art pharmacogenomics and transfer learning baselines on cell lines, patient-derived xenografts, and patients and therefore, may guide precision oncology more accurately.

show abstract

“…• The Cancer Therapeutics Response Portal (CTRPv2) [1,2] • The Genentech Cell Line Screening Initiative (gCSI) [10,11] • The Genomics of Drug Sensitivity in Cancer (GDSCv1 and GDSCv2) [3,4] We obtained these datasets in the format of PharmacoSet (PSet) which is an R-based data structure that aids in reproducible research for drug sensitivity prediction. PSets are obtained via the ORCESTRA platform (orcestra.ca) [23]. The molecular profiles (RNA-seq) were preprocessed via Kallisto 0.46.1 [50] using GENCODE v33 transcriptome as the reference and the pharmacological profiles (AAC and IC50) were preprocessed and recomputed via PharmacoGx package [49].…”

Section: Datasetsmentioning

confidence: 99%

“…We picked GDSCv1 as the competitor because it is the most common training dataset (Figure 2). GDSCv1 utilizes a different drug screening assay compared to the other datasets and for the majority of the drugs it has a smaller sample size [23]. Our hypothesis is that models trained on CTRPv2 are more generalizable because it utilizes the same assay as other datasets and also has a relatively larger sample size.…”

Section: Experimental Designmentioning

confidence: 99%

“…Due to the complexity of generating pharmacogenomics datasets, discrepancies can even exist across cell line datasets [11,[16][17][18][19][20][21][22]. However, recent efforts such as the PharmacoDB project (pharmacodb.ca) [15], the ORCESTRA platform (orcestra.ca) [23], and CellMinerCDB [24] aimed at standardizing, and integrating different preclinical pharmacogenomics datasets to improve downstream machine learning modeling. The data-rich nature of preclinical pharmacogenomics datasets has paved the way for the development of machine learning approaches to predict drug sensitivity in vitro and in vivo [25][26][27].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Drug Sensitivity Prediction From Cell Line-Based Pharmacogenomics Data: Guidelines for Developing Machine Learning Models

Noghabi

Jahangiri-Tazehkand

Hon

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

The goal of precision oncology is to tailor treatment for patients individually using the genomic profile of their tumors. Pharmacogenomics datasets such as cancer cell lines are among the most valuable resources for drug sensitivity prediction, a crucial task of precision oncology. Machine learning methods have been employed to predict drug sensitivity based on the multiple omics data available for large panels of cancer cell lines. However, there are no comprehensive guidelines on how to properly train and validate such machine learning models for drug sensitivity prediction. In this paper, we introduce a set of guidelines for different aspects of training a predictor using cell line datasets. These guidelines provide extensive analysis of the generalization of drug sensitivity predictors, and challenge many current practices in the community including the choice of training dataset and measure of drug sensitivity. Application of these guidelines in future studies will enable the development of more robust preclinical biomarkers.

show abstract

Orchestrating and sharing large multimodal data for transparent and reproducible research

Cited by 6 publications

References 69 publications

PharmacoDB 2.0 : Improving scalability and transparency ofin vitropharmacogenomics analysis

PharmacoDB 2.0 : Improving scalability and transparency ofin vitropharmacogenomics analysis

Out-of-Distribution Generalization from Labeled and Unlabeled Gene Expression Data for Drug Response Prediction

Drug Sensitivity Prediction From Cell Line-Based Pharmacogenomics Data: Guidelines for Developing Machine Learning Models

Contact Info

Product

Resources

About