A major challenge in cancer treatment is predicting clinical response to anti-cancer drugs on a personalized basis. Using a pharmacogenomics database of 1,001 cancer cell lines, we trained deep neural networks for prediction of drug response and assessed their performance on multiple clinical cohorts. We demonstrate that deep neural networks outperform the current state in machine learning frameworks. We provide a proof of concept for the use of deep neural network-based frameworks to aid precision oncology strategies.
A major challenge in cancer treatment is predicting the clinical response to anti-cancer drugs on a personalized basis. The success of such a task largely depends on the ability to develop computational resources that integrate big "omic" data into effective drug-response models. Machine learning is both an expanding and an evolving computational field that holds promise to cover such needs. Here we provide a focused overview of: 1) the various supervised and unsupervised algorithms used specifically in drug response prediction applications, 2) the strategies employed to develop these algorithms into applicable models, 3) data resources that are fed into these frameworks and 4) pitfalls and challenges to maximize model performance. In this context we also describe a novel in silico screening process, based on Association Rule Mining, for identifying genes as candidate drivers of drug response and compare it with relevant data mining frameworks, for which we generated a web application freely available at: https://compbio.nyumc.org/drugs/. This pipeline explores with high efficiency large sample-spaces, while is able to detect low frequency events and evaluate statistical significance even in the multidimensional space, presenting the results in the form of easily interpretable rules. We conclude with future prospects and challenges of applying machine learning based drug response prediction in precision medicine.
BACKGROUND: Senescence is a fundamental biological process implicated in various pathologies, including cancer. Regarding carcinogenesis, senescence signifies, at least in its initial phases, an anti-tumor response that needs to be circumvented for cancer to progress. Micro-RNAs, a subclass of regulatory, non-coding RNAs, participate in senescence regulation. At the subcellular level micro-RNAs, similar to proteins, have been shown to traffic between organelles influencing cellular behavior. The differential function of micro-RNAs relative to their subcellular localization and their role in senescence biology raises concurrent in situ analysis of coding and non-coding gene products in senescent cells as a necessity. However, technical challenges have rendered in situ co-detection unfeasible until now. METHODS: In the present report we describe a methodology that bypasses these technical limitations achieving for the first time simultaneous detection of both a micro-RNA and a protein in the biological context of cellular senescence, utilizing the new commercially available SenTraGorTM compound. The method was applied in a prototypical human non-malignant epithelial model of oncogene-induced senescence that we generated for the purposes of the study. For the characterization of this novel system, we applied a wide range of cellular and molecular techniques, as well as high-throughput analysis of the transcriptome and micro-RNAs. RESULTS: This experimental setting has three advantages that are presented and discussed: i) it covers a "gap" in the molecular carcinogenesis field, as almost all corresponding in vitro models are fibroblast-based, even though the majority of neoplasms have epithelial origin, ii) it recapitulates the precancerous and cancerous phases of epithelial tumorigenesis within a short time frame under the light of natural selection and iii) it uses as an oncogenic signal, the replication licensing factor CDC6, implicated in both DNA replication and transcription when over-expressed, a characteristic that can be exploited to monitor RNA dynamics. CONCLUSIONS: Consequently, we demonstrate that our model is optimal for studying the molecular basis of epithelial carcinogenesis shedding light on the tumor-initiating events. The latter may reveal novel molecular targets with clinical benefit. Besides, since this method can be incorporated in a wide range of low, medium or high-throughput image-based approaches, we expect it to be broadly applicable.
- MeSH
- epitelové buňky metabolismus MeSH
- genom MeSH
- jaderné proteiny metabolismus MeSH
- karcinogeneze MeSH
- kultivované buňky MeSH
- lidé MeSH
- mikro RNA metabolismus MeSH
- nádory glandulární a epitelové genetika patologie ultrastruktura MeSH
- onkogeny * MeSH
- proteiny buněčného cyklu metabolismus MeSH
- proteiny metabolismus MeSH
- stanovení celkové genové exprese MeSH
- stárnutí buněk genetika MeSH
- Check Tag
- lidé MeSH
- Publikační typ
- časopisecké články MeSH
- práce podpořená grantem MeSH
The cyclin-dependent kinase inhibitor p21(WAF1/CIP1) (p21) is a cell-cycle checkpoint effector and inducer of senescence, regulated by p53. Yet, evidence suggests that p21 could also be oncogenic, through a mechanism that has so far remained obscure. We report that a subset of atypical cancerous cells strongly expressing p21 showed proliferation features. This occurred predominantly in p53-mutant human cancers, suggesting p53-independent upregulation of p21 selectively in more aggressive tumour cells. Multifaceted phenotypic and genomic analyses of p21-inducible, p53-null, cancerous and near-normal cellular models showed that after an initial senescence-like phase, a subpopulation of p21-expressing proliferating cells emerged, featuring increased genomic instability, aggressiveness and chemoresistance. Mechanistically, sustained p21 accumulation inhibited mainly the CRL4-CDT2 ubiquitin ligase, leading to deregulated origin licensing and replication stress. Collectively, our data reveal the tumour-promoting ability of p21 through deregulation of DNA replication licensing machinery-an unorthodox role to be considered in cancer treatment, since p21 responds to various stimuli including some chemotherapy drugs.
- MeSH
- cykliny genetika metabolismus MeSH
- inhibitor p21 cyklin-dependentní kinasy genetika metabolismus MeSH
- kultivované buňky MeSH
- lidé MeSH
- nádorový supresorový protein p53 genetika metabolismus MeSH
- nádory genetika MeSH
- nestabilita genomu genetika MeSH
- replikace DNA genetika MeSH
- Check Tag
- lidé MeSH
- Publikační typ
- časopisecké články MeSH
