Mutation-Attention (MuAt): deep representation learning of somatic mutations for tumour typing and subtyping

Sanjaya, Prima; Waszak, Sebastian M.; Stegle, Oliver; Korbel, Jan O.; Pitkänen, Esa

doi:10.1101/2022.03.15.483816

Cited by 4 publications

(5 citation statements)

References 64 publications

(104 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although MuSiCal solves several problems in the standard workflow, other methods that perform specific tasks – such as joint discovery of different signature types [50], incorporation of epigenetic data [51], and utilization of clinical information [52] – are complementary and valuable additions to the signature analysis toolbox. There are also many other outstanding methodological challenges requiring future developments, which could potentially benefit from recent advances in machine learning and artificial intelligence [53].…”

Section: Discussionmentioning

confidence: 99%

Accurate and sensitive mutational signature analysis with MuSiCal

Gulhan

Ben-Isvy

et al. 2022

Preprint

View full text Add to dashboard Cite

Mutational signature analysis is a recent computational approach for interpreting somatic mutations in the genome. Its application to cancer data has enhanced our understanding of mutational forces driving tumorigenesis and demonstrated its potential to inform prognosis and treatment decisions. However, methodological challenges remain for discovering new signatures and assigning proper weights to existing signatures, thereby hindering broader clinical applications. We present MuSiCal (Mutational Signature Calculator), a rigorous analytical framework with novel algorithms that solves fundamental problems in the standard workflow. Our simulation studies demonstrate that MuSiCal outperforms state-of-the-art algorithms for both signature discovery and assignment. By reanalyzing over 2,700 cancer genomes, we provide an improved reference catalog of signatures and their assignments, discover nine indel signatures absent in the current catalog, resolve long-standing issues with the ambiguous ‘flat’ signatures, and give insights into signatures with unknown etiologies. We expect MuSiCal and the improved catalog to be a step towards establishing best practices for mutational signature analysis.

show abstract

Section: Discussionmentioning

confidence: 99%

Accurate and sensitive mutational signature analysis with MuSiCal

Gulhan

Ben-Isvy

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Machine learning is a rapidly evolving subfield of artificial intelligence, which utilises algorithms and statistical models to analyse and detect patterns in large and complex datasets to aid diagnosis. Machine learning approaches have been applied to digital pathology [38], RNA [39,40], NGS [41 & ,42] or WGS [43][44][45] results to help determine ToO (Table 3 ] developed a CUP prediction algorithm (CUPPA) based on orthogonal information of passenger and driver mutational features, regional mutational density and SNV-96 mutational profile utilising a cancer WGS database developed by the Hartwig Medical Foundation in the Netherlands. When CUPPA results were integrated in the diagnostic work-up, CUPPA could identify a primary tumour type for 49/ 72 CUP patients (68%).…”

Section: Prospective Clinical Trials Evaluating the Clinical Utility ...mentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

The evolution of molecular management of carcinoma of unknown primary

Sivakumaran,

Tothill,

Mileshkin

2024

Current Opinion in Oncology

View full text Add to dashboard Cite

Purpose of review There is significant need to improve diagnostic and therapeutic options for patients with cancer of unknown primary (CUP). In this review, we discuss the evolving landscape of molecular profiling in CUP. Recent findings Molecular profiling is becoming accepted into the diagnostic work-up of CUP patients with tumour mutation profiling now described in international CUP guidelines. Although tissue-of-origin (ToO) molecular tests utilising gene-expression and DNA methylation have existed some time, their clinical benefit remains unclear. Novel technologies utilising whole genome sequencing and machine learning algorithms are showing promise in determining ToO, however further research is required prior to clinical application. A recent international clinical trial found patients treated with molecularly-guided therapy based on comprehensive-panel DNA sequencing had improved progression-free survival compared to chemotherapy alone, confirming utility of performing genomic profiling early in the patient journey. Small phase 2 trials have demonstrated that some CUP patients are responsive to immunotherapy, but the best way to select patients for treatment is not clear. Summary Management of CUP requires a multifaceted approach incorporating clinical, histopathological, radiological and molecular sequencing results to assist with identifying the likely ToO and clinically actionable genomic alternations. Rapidly identifying a subset of CUP patients who are likely to benefit from site specific therapy, targeted therapy and/or immunotherapy will improve patient outcomes.

show abstract

“…https://doi.org/10.1038/s41551-023-01120-3 histology of origin, and these features of somatic mutations have been shown to be capable of classifying cancers 17,[20][21][22] . We were interested in seeing how a model that calculates attention or learns it owns features compared to established approaches.…”

Section: Articlementioning

confidence: 99%

Multiple-instance learning of somatic mutations for the classification of tumour type and the prediction of microsatellite status

Anaya,

Sidhom,

Mahmood

et al. 2023

Nat. Biomed. Eng

View full text Add to dashboard Cite

Large-scale genomic data are well suited to analysis by deep learning algorithms. However, for many genomic datasets, labels are at the level of the sample rather than for individual genomic measures. Machine learning models leveraging these datasets generate predictions by using statically encoded measures that are then aggregated at the sample level. Here we show that a single weakly supervised end-to-end multiple-instance-learning model with multi-headed attention can be trained to encode and aggregate the local sequence context or genomic position of somatic mutations, hence allowing for the modelling of the importance of individual measures for sample-level classification and thus providing enhanced explainability. The model solves synthetic tasks that conventional models fail at, and achieves best-in-class performance for the classification of tumour type and for predicting microsatellite status. By improving the performance of tasks that require aggregate information from genomic datasets, multiple-instance deep learning may generate biological insight.

show abstract

Mutation-Attention (MuAt): deep representation learning of somatic mutations for tumour typing and subtyping

Cited by 4 publications

References 64 publications

Accurate and sensitive mutational signature analysis with MuSiCal

Accurate and sensitive mutational signature analysis with MuSiCal

The evolution of molecular management of carcinoma of unknown primary

Multiple-instance learning of somatic mutations for the classification of tumour type and the prediction of microsatellite status

Contact Info

Product

Resources

About