Inflation of tumor mutation burden by tumor-only sequencing in under-represented groups

Asmann, Yan W.; Parikh, Kaushal; Bergsagel, P. Leif; Dong, Haidong; Adjei, Alex A.; Borad, Mitesh J.; Mansfield, Aaron S.

doi:10.1038/s41698-021-00164-5

Cited by 20 publications

(26 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There are several limitations of this application. Sources of TMB measurement error other than those derived from sampling such as variant allele fraction thresholds, distinction between germline and somatic variants (which may be particularly challenging for patients from under-represented groups [ 19 ]), and potential targeted panel bias toward hotspot mutations are not reflected in the output of the application. These additional sources of error are likely substantial when using small panels, where each incremental mutation identified has a large effect on the TMB calculation.…”

Section: Discussionmentioning

confidence: 99%

Visualization of the Effect of Assay Size on the Error Profile of Tumor Mutational Burden Measurement

Bailey

2022

Genes

View full text Add to dashboard Cite

Tumor mutational burden (TMB) refers to the number of somatic mutations in a tumor per megabase and is a biomarker for response to immune checkpoint inhibitor therapy. Immune checkpoint inhibitors are currently approved for tumors with TMB greater than or equal to 10 mutations/megabase. Many laboratories are currently reporting TMB values based upon targeted resequencing panels with limited genomic coverage. Due to sampling variation, this leads to significant uncertainty in the assay’s TMB result, particularly at relatively low TMB levels near the 10 mutation per megabase therapeutic threshold. In order to allow clinicians and laboratorians to explore this uncertainty, we built a novel web application that allows a user to view the potential error of a TMB result given the sequencing panel size. This application also allows the user to explore the effect of incorporating knowledge of a specific tumor type’s typical TMB distribution on the error profile of the TMB result.

show abstract

Section: Discussionmentioning

confidence: 99%

Visualization of the Effect of Assay Size on the Error Profile of Tumor Mutational Burden Measurement

Bailey

2022

Genes

View full text Add to dashboard Cite

show abstract

“…A major limitation in human genomics and precision medicine is that not all subpopulations are well-represented in genomic studies (Popejoy and Fullerton 2016). Human germline variant databases predominantly consist of subjects of White European ancestry, and this bias diminishes the reliability of naive tumor-only variant calling methods for Blacks relative to Whites (Asmann et al 2021;Halperin et al 2017). By integrating multiple informative features such as COSMIC and germline databases, variant allele fractions, and the local copy number ratios of known heterozygous germline SNPs, we can reduce much of this racial bias.…”

Section: Discussionmentioning

confidence: 99%

“…Tumor-only variant calling is no exception (Halperin et al 2017). A recent study observed that the inflation of TMB caused by the absence of a matched normal sample is most severe in underrepresented minorities (Asmann et al 2021).…”

Section: Impact Of Racially Biased Germline Databases In Tumor-only Variant Callingmentioning

confidence: 99%

Attentive deep learning-based tumor-only somatic mutation classifier achieves high accuracy agnostic of tissue type and capture kit

McLaughlin

Asthana

Meo

et al. 2021

Preprint

View full text Add to dashboard Cite

In precision oncology, reliable identification of tumor-specific DNA mutations requires sequencing tumor DNA and non-tumor DNA (so-called "matched normal") from the same patient. The normal sample allows researchers to distinguish acquired (somatic) and hereditary (germline) variants. The ability to distinguish somatic and germline variants facilitates estimation of tumor mutation burden (TMB), which is a recently FDA-approved pan-cancer marker for highly successful cancer immunotherapies; in tumor-only variant calling (i.e., without a matched normal), the difficulty in discriminating germline and somatic variants results in inflated and unreliable TMB estimates. We apply machine learning to the task of somatic vs germline classification in tumor-only samples using TabNet, a recently developed attentive deep learning model for tabular data that has achieved state of the art performance in multiple classification tasks (Arik and Pfister 2019). We constructed a training set for supervised classification using features derived from tumor-only variant calling and drawing somatic and germline truth-labels from an independent pipeline incorporating the patient-matched normal samples. Our trained model achieved state-of-the-art performance on two hold-out test datasets: a TCGA dataset including sarcoma, breast adenocarcinoma, and endometrial carcinoma samples (F1-score: 88.3), and a metastatic melanoma dataset, (F1-score 79.8). Concordance between matched-normal and tumor-only TMB improves from R2 = 0.006 to 0.705 with the addition of our classifier. And importantly, this approach generalizes across tumor tissue types and capture kits and has a call rate of 100%. The interpretable feature masks of the attentive deep learning model explain the reasons for misclassified variants. We reproduce the recent finding that tumor-only TMB estimates for Black patients are extremely inflated relative to that of White patients due to the racial biases of germline databases. We show that our machine learning approach appreciably reduces this racial bias in tumor-only variant-calling.

show abstract

“…TMB is defined as the total number of nonsynonymous mutations of coding regions of a tumor genome per mega-base (Mb) of a tumor genome ( 19 ). Mutations from the 39 tumor samples were clustered into subclones using PyClone-0.13.0 based on SNVs ( 20 ), and copy number variations were analyzed using FACETS.…”

Section: Methodsmentioning

confidence: 99%

RBM10 Deficiency Is Associated With Increased Immune Activity in Lung Adenocarcinoma

Liu

Wang²,

Wang³

et al. 2021

Front. Oncol.

View full text Add to dashboard Cite

IntroductionRBM10 is one of the frequently mutated genes in lung adenocarcinoma (LUAD). Previous studies have confirmed that RBM10 could suppress the disease progression and cell proliferation in LUAD, but its loss-of-function mutations are more frequent in early-stage disease and decrease with the advancement of the clinical stage. This is contradictory to its role of tumor suppressor. Here, we conducted a systematic analysis to elucidate whether there was other potential biological significance of RBM10 deficiency during the progression of LUAD.Materials and MethodsThe whole exome sequencing data of 39 tumor samples from early-stage LUADs (GGN cohort) and genomic and transcriptome data of the Cancer Genome Atlas (TCGA) LUAD cohort (TCGA_LUAD cohort) and a Chinese LUAD cohort (CHOICE_ADC cohort) were first obtained. Systematic bioinformatic analyses were then conducted to determine gene expression signature, immune infiltration levels and predicted immunotherapy response. Immunohistochemistry (IHC) was also conducted to validate the result of immune infiltration.ResultsThe mutation rate of RBM10 was significantly higher in the GGN cohort than that in the TCGA_LUAD and CHOICE_ADC cohorts. In both TCGA_LUAD and CHOICE_ADC cohorts, multiple immune related pathways were markedly enriched in RBM10 deficient group. Further analyses showed that tumors with RBM10 mutations displayed higher TMB, and LUADs with RBM10 deficiency also showed higher HLA expression levels, including many HLA class I and II molecules. Additionally, many immune cells, including myeloid dendritic cells, macrophages, neutrophils and CD8+T cells, showed higher infiltration levels in LUADs with RBM10 deficiency. Finally, some immune checkpoint molecules, such as PD-L1 and TIM-3, were highly expressed in RBM10 deficient population and the predicted immunotherapy response was calculated through TIDE algorithm, showing that IFNG expression, MSI score and CD8 expression were higher in RBM10 deficient group, while MDSC and M2 macrophage were lower in RBM10 deficient group.ConclusionOur study demonstrates that RBM10 deficient LUADs show higher HLA expression and immune cell infiltration, and some immune checkpoint molecules are also highly expressed. In brief, RBM10 deficiency could enhance anti-tumor immunity in LUAD.

show abstract

Inflation of tumor mutation burden by tumor-only sequencing in under-represented groups

Cited by 20 publications

References 14 publications

Visualization of the Effect of Assay Size on the Error Profile of Tumor Mutational Burden Measurement

Visualization of the Effect of Assay Size on the Error Profile of Tumor Mutational Burden Measurement

Attentive deep learning-based tumor-only somatic mutation classifier achieves high accuracy agnostic of tissue type and capture kit

RBM10 Deficiency Is Associated With Increased Immune Activity in Lung Adenocarcinoma

Contact Info

Product

Resources

About