A deep learning system accurately classifies primary and metastatic cancers using passenger mutation patterns.
Ontology highlight
ABSTRACT:
SUBMITTER: Jiao W
PROVIDER: S-EPMC7002586 | biostudies-literature | 2020 Feb
REPOSITORIES: biostudies-literature
Publications
A deep learning system accurately classifies primary and metastatic cancers using passenger mutation patterns.
Jiao Wei W Atwal Gurnit G Polak Paz P Karlic Rosa R Cuppen Edwin E Danyi Alexandra A de Ridder Jeroen J van Herpen Carla C Lolkema Martijn P MP Steeghs Neeltje N Getz Gad G Morris Quaid D QD Stein Lincoln D LD
Nature communications 20200205 1
In cancer, the primary tumour's organ of origin and histopathology are the strongest determinants of its clinical behaviour, but in 3% of cases a patient presents with a metastatic tumour and no obvious primary. Here, as part of the ICGC/TCGA Pan-Cancer Analysis of Whole Genomes (PCAWG) Consortium, we train a deep learning classifier to predict cancer type based on patterns of somatic passenger mutations detected in whole genome sequencing (WGS) of 2606 tumours representing 24 common cancer type ...[more]
Similar Datasets
Project description:Cardiovascular diseases (CVDs) have become the number 1 threat to human health. Their numerous complications mean that many countries remain unable to prevent the rapid growth of such diseases, although significant health resources have been invested toward their prevention and management. Electrocardiogram (ECG) is the most important non-invasive physiological signal for CVD screening and diagnosis. For exploring the heartbeat event classification model using single- or multiple-lead ECG signals, we proposed a novel deep learning algorithm and conducted a systemic comparison based on the different methods and databases. This new approach aims to improve accuracy and reduce training time by combining the convolutional neural network (CNN) with the bidirectional long short-term memory (BiLSTM). To our knowledge, this approach has not been investigated to date. In this study, Database I with single-lead ECG and Database II with 12-lead ECG were used to explore a practical and viable heartbeat event classification model. An evolutionary neural system approach (Method I) and a deep learning approach (Method II) that combines CNN with BiLSTM network were compared and evaluated in processing heartbeat event classification. Overall, Method I achieved slightly better performance than Method II. However, Method I took, on average, 28.3 h to train the model, whereas Method II needed only 1 h. Method II achieved an accuracy of 80, 82.6, and 85% compared with the China Physiological Signal Challenge 2018, PhysioNet Challenge 2017, and Massachusetts Institute of Technology-Beth Israel Hospital (MIT-BIH) Arrhythmia datasets, respectively. These results are impressive compared with the performance of state-of-the-art algorithms used for the same purpose.
| S-EPMC7566908 | biostudies-literature
Project description:The mutational profiles of primary colorectal cancers (CRCs) and corresponding ovarian metastases were compared. Using a custom-made next generation sequencing panel, 115 cancer-driving genes were analyzed in a cohort of 26 primary CRCs and 30 matching ovarian metastases (four with bilateral metastases). To obtain a complete overview of the mutational profile, low thresholds were used in bioinformatics analysis to prevent low frequency passenger mutations from being filtered out. A subset of variants was validated using Sanger and/or hydrolysis probe assays. The mutational landscape of CRC that metastasized to the ovary was not strikingly different from CRC in consecutive series. When comparing primary CRCs and their matching ovarian metastases, there was considerable overlap in the mutations of early affected genes. A subset of mutations demonstrated less overlap, presumably being passenger mutations. In particular, primary CRCs showed a substantially high number of passenger mutations. We also compared the primary CRCs and matching metastases for stratifying variants of six genes (KRAS, NRAS, BRAF, FBXW7, PTEN and PIK3CA) that select for established (EGFR directed) or future targeted therapies. In a total of 31 variants 12 were not found in either of the two locations. Tumours thus differed in the number of discordant variants between the primary tumours and matching metastases. Half of these discordant variants were definitive class 4/5 pathogenic variants. However, in terms of temporal heterogeneity, no clear relationship was observed between the number of discordant variants and the time interval between primary CRCs and the detection of ovarian metastases. This suggests that dormant metastases may be present from the early days of the primary tumours.
| S-EPMC4958737 | biostudies-literature
Project description:Determining the cancer type and molecular subtype has important clinical implications. The primary site is however unknown for some malignancies discovered in the metastatic stage. Moreover liquid biopsies may be used to screen for tumoral DNA, which upon detection needs to be assigned to a site-of-origin. Classifiers based on genomic features are a promising approach to prioritize the tumor anatomical site, type and subtype. We examined the predictive ability of causal (driver) somatic mutations in this task, comparing it against global patterns of non-selected (passenger) mutations, including features based on regional mutation density (RMD). In the task of distinguishing 18 cancer types, the driver mutations-mutated oncogenes or tumor suppressors, pathways and hotspots-classified 36% of the patients to the correct cancer type. In contrast, the features based on passenger mutations did so at 92% accuracy, with similar contribution from the RMD and the trinucleotide mutation spectra. The RMD and the spectra covered distinct sets of patients with predictions. In particular, introducing the RMD features into a combined classification model increased the fraction of diagnosed patients by 50 percentage points (at 20% FDR). Furthermore, RMD was able to discriminate molecular subtypes and/or anatomical site of six major cancers. The advantage of passenger mutations was upheld under high rates of false negative mutation calls and with exome sequencing, even though overall accuracy decreased. We suggest whole genome sequencing is valuable for classifying tumors because it captures global patterns emanating from mutational processes, which are informative of the underlying tumor biology.
| S-EPMC6483366 | biostudies-literature
Project description:BackgroundIdentifying breast cancer patients with DNA repair pathway-related germline pathogenic variants (GPVs) is important for effectively employing systemic treatment strategies and risk-reducing interventions. However, current criteria and risk prediction models for prioritizing genetic testing among breast cancer patients do not meet the demands of clinical practice due to insufficient accuracy.MethodsThe study population comprised 3041 breast cancer patients enrolled from seven hospitals between October 2017 and 11 August 2019, who underwent germline genetic testing of 50 cancer predisposition genes (CPGs). Associations among GPVs in different CPGs and endophenotypes were evaluated using a case-control analysis. A phenotype-based GPV risk prediction model named DNA-repair Associated Breast Cancer (DrABC) was developed based on hierarchical neural network architecture and validated in an independent multicenter cohort. The predictive performance of DrABC was compared with currently used models including BRCAPRO, BOADICEA, Myriad, PENN II, and the NCCN criteria.ResultsIn total, 332 (11.3%) patients harbored GPVs in CPGs, including 134 (4.6%) in BRCA2, 131 (4.5%) in BRCA1, 33 (1.1%) in PALB2, and 37 (1.3%) in other CPGs. GPVs in CPGs were associated with distinct endophenotypes including the age at diagnosis, cancer history, family cancer history, and pathological characteristics. We developed a DrABC model to predict the risk of GPV carrier status in BRCA1/2 and other important CPGs. In predicting GPVs in BRCA1/2, the performance of DrABC (AUC = 0.79 [95% CI, 0.74-0.85], sensitivity = 82.1%, specificity = 63.1% in the independent validation cohort) was better than that of previous models (AUC range = 0.57-0.70). In predicting GPVs in any CPG, DrABC (AUC = 0.74 [95% CI, 0.69-0.79], sensitivity = 83.8%, specificity = 51.3% in the independent validation cohort) was also superior to previous models in their current versions (AUC range = 0.55-0.65). After training these previous models with the Chinese-specific dataset, DrABC still outperformed all other methods except for BOADICEA, which was the only previous model with the inclusion of pathological features. The DrABC model also showed higher sensitivity and specificity than the NCCN criteria in the multi-center validation cohort (83.8% and 51.3% vs. 78.8% and 31.2%, respectively, in predicting GPVs in any CPG). The DrABC model implementation is available online at http://gifts.bio-data.cn/ .ConclusionsBy considering the distinct endophenotypes associated with different CPGs in breast cancer patients, a phenotype-driven prediction model based on hierarchical neural network architecture was created for identification of hereditary breast cancer. The model achieved superior performance in identifying GPV carriers among Chinese breast cancer patients.
| S-EPMC8876403 | biostudies-literature
Project description:The activPAL (AP) monitor is well established for distinguishing sitting, standing, and stepping time. However, its validity in predicting time in physical activity intensity categories in a free-living environment has not been determined.PurposeThis study aimed to determine the validity of the AP in estimating time spent in sedentary, light, and moderate-to-vigorous physical activity (MVPA) in a free-living environment.MethodsThirteen participants (mean ± SD age = 24.8 ± 5.2 yr, BMI = 23.8 ± 1.9 kg·m) were directly observed for three 10-h periods wearing an AP. A custom R program was developed and used to summarize detailed active and sedentary behavior variables from the AP. AP estimates were compared with direct observation.ResultsThe AP accurately and precisely estimated time in activity intensity categories (bias [95% confidence interval]; sedentary = 0.8 min [-2.9 to 4.5], light = 1.7 min [2.2-5.7], and -2.6 min [-5.8 to 0.7]). The overall accuracy rate for time in intensity categories was 96.2%. The AP also accurately estimated guideline minutes, guideline bouts, prolonged sitting minutes, and prolonged sitting bouts.ConclusionThe AP can be used to accurately capture individualized estimates of active and sedentary behavior variables in free-living settings.
| S-EPMC5469371 | biostudies-literature
Project description:The tumor suppressor gene, TP53, has the highest rate of mutation among all genes in human cancer. This transcription factor plays an essential role in the regulation of many cellular processes. Mutations in TP53 result in loss of wild-type p53 function in a dominant negative manner. Although TP53 is a well-studied gene, the transcriptome modifications caused by the mutations in this gene have not yet been explored in a pan-cancer study using both primary and metastatic samples. In this work, we used a random forest model to stratify tumor samples based on TP53 mutational status and detected a p53 transcriptional signature. We hypothesize that the existence of this transcriptional signature is due to the loss of wild-type p53 function and is universal across primary and metastatic tumors as well as different tumor types. Additionally, we showed that the algorithm successfully detected this signature in samples with apparent silent mutations that affect correct mRNA splicing. Furthermore, we observed that most of the highly ranked genes contributing to the classification extracted from the random forest have known associations with p53 within the literature. We suggest that other genes found in this list including GPSM2, OR4N2, CTSL2, SPERT, and RPE65 protein coding genes have yet undiscovered linkages to p53 function. Our analysis of time on different therapies also revealed that this signature is more effective than the recorded TP53 status in detecting patients who can benefit from platinum therapies and taxanes. Our findings delineate a p53 transcriptional signature, expand the knowledge of p53 biology and further identify genes important in p53 related pathways.
| S-EPMC9483853 | biostudies-literature
Project description:Plasmids are mobile genetic elements that play a key role in microbial ecology and evolution by mediating horizontal transfer of important genes, such as antimicrobial resistance genes. Many microbial genomes have been sequenced by short read sequencers and have resulted in a mix of contigs that derive from plasmids or chromosomes. New tools that accurately identify plasmids are needed to elucidate new plasmid-borne genes of high biological importance. We have developed Deeplasmid, a deep learning tool for distinguishing plasmids from bacterial chromosomes based on the DNA sequence and its encoded biological data. It requires as input only assembled sequences generated by any sequencing platform and assembly algorithm and its runtime scales linearly with the number of assembled sequences. Deeplasmid achieves an AUC-ROC of over 89%, and it was more accurate than five other plasmid classification methods. Finally, as a proof of concept, we used Deeplasmid to predict new plasmids in the fish pathogen Yersinia ruckeri ATCC 29473 that has no annotated plasmids. Deeplasmid predicted with high reliability that a long assembled contig is part of a plasmid. Using long read sequencing we indeed validated the existence of a 102 kb long plasmid, demonstrating Deeplasmid's ability to detect novel plasmids.
| S-EPMC8860608 | biostudies-literature
Project description:Uterine leiomyomas are benign smooth muscle tumors occurring in 70% of women of reproductive age. The majority of leiomyomas harbor one of three well-established genetic changes: a hotspot mutation in MED12, overexpression of HMGA2, or biallelic loss of FH. The majority of studies have classified leiomyomas by complex and costly methods, such as whole-genome sequencing, or by combining multiple traditional methods, such as immunohistochemistry and Sanger sequencing. The type of specimens and the amount of resources available often determine the choice. A more universal, cost-effective, and scalable method for classifying leiomyomas is needed. The aim of this study was to evaluate whether RNA sequencing can accurately classify formalin-fixed paraffin-embedded (FFPE) leiomyomas. We performed 3'RNA sequencing with 44 leiomyoma and 5 myometrium FFPE samples, revealing that the samples clustered according to the mutation status of MED12, HMGA2, and FH. Furthermore, we confirmed each subtype in a publicly available fresh frozen dataset. These results indicate that a targeted 3'RNA sequencing panel could serve as a cost-effective and robust tool for stratifying both fresh frozen and FFPE leiomyomas. This study also highlights 3'RNA sequencing as a promising method for studying the abundance of unexploited tissue material that is routinely stored in hospital archives.
| S-EPMC7766537 | biostudies-literature
Project description:Accurately predicting essential genes using computational methods can greatly reduce the effort in finding them via wet experiments at both time and resource scales, and further accelerate the process of drug discovery. Several computational methods have been proposed for predicting essential genes in model organisms by integrating multiple biological data sources either via centrality measures or machine learning based methods. However, the methods aiming to predict human essential genes are still limited and the performance still need improve. In addition, most of the machine learning based essential gene prediction methods are lack of skills to handle the imbalanced learning issue inherent in the essential gene prediction problem, which might be one factor affecting their performance. We propose a deep learning based method, DeepHE, to predict human essential genes by integrating features derived from sequence data and protein-protein interaction (PPI) network. A deep learning based network embedding method is utilized to automatically learn features from PPI network. In addition, 89 sequence features were derived from DNA sequence and protein sequence for each gene. These two types of features are integrated to train a multilayer neural network. A cost-sensitive technique is used to address the imbalanced learning problem when training the deep neural network. The experimental results for predicting human essential genes show that our proposed method, DeepHE, can accurately predict human gene essentiality with an average performance of AUC higher than 94%, the area under precision-recall curve (AP) higher than 90%, and the accuracy higher than 90%. We also compare DeepHE with several widely used traditional machine learning models (SVM, Naïve Bayes, Random Forest, and Adaboost) using the same features and utilizing the same cost-sensitive technique to against the imbalanced learning issue. The experimental results show that DeepHE significantly outperforms the compared machine learning models. We have demonstrated that human essential genes can be accurately predicted by designing effective machine learning algorithm and integrating representative features captured from available biological data. The proposed deep learning framework is effective for such task.
| S-EPMC7521708 | biostudies-literature
Project description:Reconstructing the evolutionary history of metastases is critical for understanding their basic biological principles and has profound clinical implications. Genome-wide sequencing data has enabled modern phylogenomic methods to accurately dissect subclones and their phylogenies from noisy and impure bulk tumour samples at unprecedented depth. However, existing methods are not designed to infer metastatic seeding patterns. Here we develop a tool, called Treeomics, to reconstruct the phylogeny of metastases and map subclones to their anatomic locations. Treeomics infers comprehensive seeding patterns for pancreatic, ovarian, and prostate cancers. Moreover, Treeomics correctly disambiguates true seeding patterns from sequencing artifacts; 7% of variants were misclassified by conventional statistical methods. These artifacts can skew phylogenies by creating illusory tumour heterogeneity among distinct samples. In silico benchmarking on simulated tumour phylogenies across a wide range of sample purities (15-95%) and sequencing depths (25-800 × ) demonstrates the accuracy of Treeomics compared with existing methods.
| S-EPMC5290319 | biostudies-literature