Deep convolutional neural networks for accurate somatic mutation detection

Sahraeian, Sayed Mohammad Ebrahim; Liu, Ruolin; Lau, Bayo; Mohiyuddin, Marghoob; Lam, Hugo Y. K.

doi:10.1101/393801

Cited by 21 publications

(31 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We used several different training models in our analysis. First, we used the already available model published recently 11 which was trained using in silico spike-ins from the DREAM Challenge Stage 3 dataset 13 . Despite the large discrepancy between the sample types, sequencing platforms, coverages, spike-in mutation frequencies, and heterogeneity of the samples used to train the DREAM3 model, this model outperformed other conventional techniques across the real cancer datasets of diverse characteristics by more than ∼4% by the mean F1-score averaged across different samples for both SNVs and INDELs ( Figure 1a ).…”

Section: Resultsmentioning

confidence: 99%

“…As the baseline WGS model we employed the DREAM3 model, developed recently 11 by training on ICGC-TCGA DREAM Challenge Stage 3 data 14 . The Stage 3 dataset consists of a normal sample and a tumor sample constructed by computationally spiking 7,903 SNVs and 7,604 INDELs mutations into a healthy genome of the same normal sample with three different AFs of 50%, 33%, and 20% to create synthetic but realistic tumoral normal pairs.…”

Section: Methodsmentioning

confidence: 99%

“…The Stage 3 dataset consists of a normal sample and a tumor sample constructed by computationally spiking 7,903 SNVs and 7,604 INDELs mutations into a healthy genome of the same normal sample with three different AFs of 50%, 33%, and 20% to create synthetic but realistic tumoral normal pairs. An additional ∼95K SNVs and ∼95K INDELs were spike-in into the tumor sample 11 using BAMSurgeon 17 with similar AF distributions to the original DREAM data for better network training. This model was trained by combining training data (in 50% of the genome) from five different tumor-normal purity setting of 100T:100N, 50T:100N, 70T:95N, 50T:95N, and 25T:95N 11 .…”

Section: Methodsmentioning

confidence: 99%

“…Accurate somatic mutation detection enables precise diagnosis, prognosis, and treatment of cancer patients 1 . Several tools have been developed to identify somatic mutations from next-generation sequencing technology 2–11 . In general, conventional techniques pinpoint somatic mutations from background noise, germline variants, and/or cross-contaminations through various statistical/algorithmic modeling approaches employing a set of hand-crafted sequencing features extracted in the tumor-normal paired sequencing data 2–10 .…”

Section: Introductionmentioning

confidence: 99%

“…Recently, a deep learning based somatic mutation detection approach, called NeuSomatic, has been proposed that uses a convolutional neural network (CNN) to learn feature representations directly from raw data 11 . NeuSomatic uses a novel summarization of tumor/normal alignment information as a set of input matrices that can be adopted to efficiently train models that learn how to effectively differentiate true somatic mutations from artifacts.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Robust Cancer Mutation Detection with Deep Learning Models Derived from Tumor-Normal Sequencing Data

Sahraeian

Fang

Mohiyuddin

et al. 2019

Preprint

Self Cite

View full text Add to dashboard Cite

Accurate detection of somatic mutations is challenging but critical to the understanding of cancer formation, progression, and treatment. We recently proposed NeuSomatic, the first deep convolutional neural network based somatic mutation detection approach and demonstrated performance advantages on in silico data. In this study, we used the first comprehensive and well-characterized somatic reference samples from the SEQC-II consortium to investigate best practices for utilizing deep learning framework in cancer mutation detection. Using the high-confidence somatic mutations established for these reference samples by the consortium, we identified strategies for building robust models on multiple datasets derived from samples representing real scenarios. The proposed strategies achieved high robustness across multiple sequencing technologies such as WGS, WES, AmpliSeq target sequencing for fresh and FFPE DNA input, varying tumor/normal purities, and different coverages (ranging from 10× - 2000×). NeuSomatic significantly outperformed conventional detection approaches in general, as well as in challenging situations such as low coverage, low mutation frequency, DNA damage, and difficult genomic regions.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Robust Cancer Mutation Detection with Deep Learning Models Derived from Tumor-Normal Sequencing Data

Sahraeian

Fang

Mohiyuddin

et al. 2019

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

Accurate and Flexible Bayesian Mutation Call from Multi-regional Tumor Samples

Moriyama

Imoto

Miyano

et al. 2019

Mathematical and Computational Oncology

View full text Add to dashboard Cite

We propose a Bayesian method termed MultiMuC for accurate detection of somatic mutations (mutation call) from multi-regional tumor sequence data sets. To improve detection performance, our method is based on the assumption of mutation sharing: if we can predict at least one tumor region has the mutation, then we can be more confident to detect a mutation in more tumor regions by lowering the original threshold of detection. We find two drawbacks in existing methods for leveraging the assumption of mutation sharing. First, existing methods do not consider the probability of the "No-TP (True Positive)" case: we could expect mutation candidates in multiple regions, but actually, no true mutations exist. Second, existing methods cannot leverage scores from other state-of-the-art mutation calling methods for a single-regional tumor. We overcome the first drawback through evaluation of the probability of the No-TP case. Next, we solve the second drawback by the idea of Bayes-factor-based model construction that enables flexible integration of probability-based mutation call scores as building blocks of a Bayesian statistical model. We empirically evaluate that our method steadily improves results from mutation calling methods for a singleregional tumor, e.g., Strelka2 and NeuSomatic, and outperforms existing methods for multi-regional tumors through a real-data-based simulation study. Our implementation of MultiMuC is available at https://github. com/takumorizo/MultiMuC.

show abstract

Artificial intelligence and machine learning in precision and genomic medicine

Quazi

2022

Med Oncol

115

View full text Add to dashboard Cite

The advancement of precision medicine in medical care has led behind the conventional symptom-driven treatment process by allowing early risk prediction of disease through improved diagnostics and customization of more effective treatments. It is necessary to scrutinize overall patient data alongside broad factors to observe and differentiate between ill and relatively healthy people to take the most appropriate path toward precision medicine, resulting in an improved vision of biological indicators that can signal health changes. Precision and genomic medicine combined with artificial intelligence have the potential to improve patient healthcare. Patients with less common therapeutic responses or unique healthcare demands are using genomic medicine technologies. AI provides insights through advanced computation and inference, enabling the system to reason and learn while enhancing physician decision making. Many cell characteristics, including gene up-regulation, proteins binding to nucleic acids, and splicing, can be measured at high throughput and used as training objectives for predictive models. Researchers can create a new era of effective genomic medicine with the improved availability of a broad range of datasets and modern computer techniques such as machine learning. This review article has elucidated the contributions of ML algorithms in precision and genome medicine.

show abstract

Deep convolutional neural networks for accurate somatic mutation detection

Cited by 21 publications

References 34 publications

Robust Cancer Mutation Detection with Deep Learning Models Derived from Tumor-Normal Sequencing Data

Robust Cancer Mutation Detection with Deep Learning Models Derived from Tumor-Normal Sequencing Data

Accurate and Flexible Bayesian Mutation Call from Multi-regional Tumor Samples

Artificial intelligence and machine learning in precision and genomic medicine

Contact Info

Product

Resources

About