Biology subjects

Parmigiani, G.

Publications and source records attributed to Parmigiani, G..

13 recordsLinked to original sources

bioRxiv2026

Pan-Cancer Genomic Scars of Alternative End Joining and Single-Strand Annealing

DNA double-strand breaks (DSBs) are among the most cytotoxic lesions encountered by cells and represent a major source of genome instability in cancer. The preferred pathway to repair resected DSBs is high-fidelity homologous recombination (HR), but error-prone annealing-dependent pathways such as single-strand annealing (SSA) and alternative end joining (Alt-EJ) generally act as backup repair mechanisms in HR-deficient contexts. However, the extent to which these pathways are engaged across tumor types and how their activity is coupled with HR deficiency remains unclear. In this work, we systematically characterize scars from SSA and Alt-EJ across 2,157 whole-genome-sequenced tumors spanning 17 cancer types, identifying 832 SSA-like and 37,359 Alt-EJ-like deletions in total. We find that Alt-EJ is the predominant backup repair pathway in HR-deficient tumors compared with SSA; however, prostate adenocarcinoma and hepatocellular carcinoma exhibit elevated SSA-like deletion burdens despite low HR-deficiency scores. Moreover, our genome-wide analysis reveals that SSA-like deletions preferentially occur in SINE-rich regions and exhibit pronounced enrichment near transcription start sites in HR-proficient lymphoid lineage tumors. Our results show that SSA- and Alt-EJ-associated genomic scars are not confined to HR-deficient tumors, but are shaped by local genomic and transcriptional context, capturing distinct dimensions of DSB repair activity beyond HR deficiency alone.

genomics↗

bioRxiv2023

Bayesian Multi-Study Non-Negative Matrix Factorization for Mutational Signatures

AO_SCPLOWBSTRACTC_SCPLOWMutational signatures shed insight into the range of mutational processes giving rise to tumors and allow a better understanding of cancer origin. They are typically identified from high-throughput sequencing data of cancer genomes using non-negative matrix factorization (NMF), and many such techniques have been developed towards this aim. However, it is often of particular interest to compare mutational signatures across multiple conditions, e.g. to understand which signatures are present across different treatments, or to identify signatures that are shared or specific across cancer types. Existing techniques within the NMF context only allow decomposition within a single dataset, so that integrating results across multiple conditions requires running separate analyses on each dataset, followed by subjective and manual comparisons of the identified signatures. To address this issue, we propose a Bayesian multi-study NMF method that jointly decomposes multiple studies or conditions to identify signatures that are common, specific, or partially shared by any subset. We propose two models: a "discovery-only" model that estimates de novo signatures in a completely unsupervised manner, and a "recovery-discovery" model that builds informative priors from previously known signatures to both update the estimates of these signatures and identify any novel signatures. We then further extend these models to estimate the effects of sample-level covariates on the exposures to each signature, enforcing sparsity through a non-local spike-and-slab prior. We demonstrate our approach on a range of simulations, and apply our method to colorectal cancer samples to show its utility.

bioinformatics↗

bioRxiv2023

Variant-specific Mendelian Risk Prediction Model

Many pathogenic sequence variants (PSVs) have been associated with increased risk of cancers. Mendelian risk prediction models use Mendelian laws of inheritance to predict the probability of having a PSV based on family history, as well as specified PSV frequency and penetrance (agespecific probability of developing cancer given genotype). Most existing models assume penetrance is the same for any PSVs in a certain gene. However, for some genes (for example, BRCA1/2), cancer risk does vary by PSV. We propose an extension of Mendelian risk prediction models to relax the assumption that risk is the same for any PSVs in a certain gene by incorporating variant-specific penetrances and illustrating these extensions on two existing Mendelian risk prediction models, BRCAPRO and PanelPRO. Our proposed BRCAPRO-variant and PanelPRO-variant models incorporate variant-specific BRCA1/2 PSVs through the region classifications. Due to the sparsity of the variant information we classify BRCA1/2 PSVs into three regions; the breast cancer clustering region (BCCR), the ovarian cancer clustering region (OCCR), and an other region. Simulations were conducted to evaluate the performance of the proposed BRCAPRO-variant model compared to the existing BRCAPRO model which assumes the penetrance is the same for any PSVs in BRCA1 (and respectively BRCA2). Simulation results showed that the BRCAPRO-variant model was well calibrated to predict region-specific BRCA1/2 carrier status with high discrimination and accuracy on the region-specific level. In addition, we showed that the BRCAPRO-variant model achieved performance gains over the existing risk prediction models in terms of calibration without loss in discrimination and accuracy. We also evaluated the performance of the two proposed models, BRCAPRO-variant and PanelPRO-variant, on a cohort of 1,961 families from the Cancer Genetics Network (CGN). We showed that our proposed models provide region-specific PSV carrier probabilities with high accuracy, while the calibration, discrimination and accuracy of gene-specific PSV carrier probabilities were comparable to the existing gene-specific models. As more variant-specific PSV penetrances become available, we have shown that Mendelian risk prediction models can be extended to integrate the additional information, providing precise variant or region-specific PSV carrier probabilities and improving future cancer risk predictions.

genetics↗

bioRxiv2021

Extent, impact, and mitigation of batch effects in tumor biomarker studies using tissue microarrays

Tissue microarrays (TMAs) have been used in thousands of cancer biomarker studies. To what extent batch effects, measurement error in biomarker levels between slides, affects TMA-based studies has not been assessed systematically. We evaluated 20 protein biomarkers on 14 TMAs with prospectively collected tumor tissue from 1,448 primary prostate cancers. In half of the biomarkers, more than 10% of biomarker variance was attributable to between-TMA differences (range, 1-48%). We implemented different methods to mitigate batch effects (R package batchtma), tested in plasmode simulation. Biomarker levels were more similar between mitigation approaches compared to uncorrected values. For some biomarkers, associations with clinical features changed substantially after addressing batch effects. Batch effects and resulting bias are not an error of an individual study but an inherent feature of TMA-based protein biomarker studies. They always need to be considered during study design and addressed analytically in studies using more than one TMA.

pathology↗

bioRxiv2020

ComBat-Seq: batch effect adjustment for RNA-Seq count data

The benefit of integrating batches of genomic data to increase statistical power in differential expression is often hindered by batch effects, or unwanted variation in data caused by differences in technical factors across batches. It is therefore critical to effectively address batch effects in genomic data. Many existing methods for batch effect adjustment assume continuous, bell-shaped Gaussian distributions for data. However in RNA-Seq studies where data are skewed, over-dispersed counts, this assumption is not appropriate and may lead to erroneous results. Negative binomial regression models have been used to better capture the properties of counts. We developed a batch correction method, ComBat-Seq, using negative binomial regression. ComBat-Seq retains the integer nature of count data in RNA-Seq studies, making the batch adjusted data compatible with common differential expression software packages that require integer counts. We show in realistic simulations that the ComBat-Seq adjusted data result in better statistical power and control of false positives in differential expression, compared to data adjusted by the other available methods. We further demonstrated in a real data example where ComBat-Seq successfully removes batch effects and recovers the biological signal in the data.

bioinformatics↗

bioRxiv2019

Covariate-Profile Similarity Weighting and Bagging Studies with the Study Strap: Multi-Study Learning for Human Neurochemical Sensing

We propose the "study strap ensemble," which combines advantages of two common approaches to fitting prediction models when multiple training datasets ("studies") are available: pooling studies and fitting one model versus averaging predictions from multiple models each fit to individual studies. The study strap ensemble fits models to bootstrapped datasets, or "pseudo-studies." These are generated by resampling from multiple studies with a hierarchical resampling scheme that generalizes the randomized cluster bootstrap. The study strap is controlled by a tuning parameter that determines the proportion of observations to draw from each study. When the parameter is set to its lowest value, each pseudo-study is resampled from only a single study. When it is high, the study strap ignores the multi-study structure and generates pseudo-studies by merging the datasets and drawing observations like a standard bootstrap. We empirically show the optimal tuning value often lies in between, and prove that special cases of the study strap draw the merged dataset and the set of original studies as pseudo-studies. We extend the study strap approach with an ensemble weighting scheme that utilizes information in the distribution of the covariates of the test dataset. Our work is motivated by neuroscience experiments using real-time neurochemical sensing during awake behavior in humans. Current techniques to perform this kind of research require measurements from an electrode placed in the brain during awake neurosurgery and rely on prediction models to estimate neurotransmitter concentrations from the electrical measurements recorded by the electrode. These models are trained by combining multiple datasets that are collected in vitro under heterogeneous conditions in order to promote accuracy of the models when applied to data collected in the brain. A prevailing challenge is deciding how to combine studies or ensemble models trained on different studies to enhance model generalizability. Our methods produce marked improvements in simulations and in this application. All methods are available in the studyStrap CRAN package.

neuroscience↗

bioRxiv2019

Robustifying genomic classifiers to batch effects via ensemble learning

Genomic data are often produced in batches due to practical restrictions, which may lead to unwanted variation in data caused by discrepancies across processing batches. Such \"batch effects\" often have negative impact on downstream biological analysis and need careful consideration. In practice, batch effects are usually addressed by specifically designed software, which merge the data from different batches, then estimate batch effects and remove them from the data. Here we focus on classification and prediction problems, and propose a different strategy based on ensemble learning. We first develop prediction models within each batch, then integrate them through ensemble weighting methods. In contrast to the typical approach of removing batch effects from the merged data, our method integrates predictions rather than data. We provide a systematic comparison between these two strategies, using studies targeting diverse populations infected with tuberculosis. In one study, we simulated increasing levels of heterogeneity across random subsets of the study, which we treat as simulated batches. We then use the two methods to develop a genomic classifier for the binary indicator of disease status. We evaluate the accuracy of prediction in another independent study targeting a different population cohort. We observed a turning point in the level of heterogeneity, after which our strategy of integrating predictions yields better discrimination in independent validation than the traditional method of integrating the data. These observations provide practical guidelines for handling batch effects in the development and evaluation of genomic classifiers.

bioinformatics↗

bioRxiv2019

Tree-weighting for multi-study ensemble learners

Multi-study learning uses multiple training studies, separately trains classifiers on individual studies, and then forms ensembles with weights rewarding members with better cross-study prediction ability. This article considers novel weighting approaches for constructing tree-based ensemble learners in this setting. Using Random Forests as a single-study learner, we perform a comparison of either weighting each forest to form the ensemble, or extracting the individual trees trained by each Random Forest and weighting them directly. We consider weighting approaches that reward cross-study replicability within the training set. We find that incorporating multiple layers of ensembling in the training process increases the robustness of the resulting predictor. Furthermore, we explore the mechanisms by which the ensembling weights correspond to the internal structure of trees to shed light on the important features in determining the relationship between the Random Forests algorithm and the true outcome model. Finally, we apply our approach to genomic datasets and show that our method improves upon the basic multi-study learning paradigm.

bioinformatics↗

bioRxiv2019

Multi-omic analysis supports a developmental hierarchy of molecular subtypes in high-grade serous ovarian carcinoma

Multiple studies have identified transcriptome subtypes of high-grade serous ovarian carcinoma (HGSOC), but these have yet to impact clinical practice. Interpretation and translation of HGSOC subtypes are complicated by tumor evolution and polyclonality accompanied by accumulation of somatic aberrations, varying cell type admixtures, and different tissues of origin. The chronology of HGSOC subtype evolution was examined in the context of these factors by a novel integrative analysis of bulk absolute somatic copy number analysis and gene expression in The Cancer Genome Atlas, complemented by single-cell RNA-seq analysis of six independent tumors. The approach was validated by contrast to soft-tissue sarcoma. Genomic lesions associated with HGSOC subtypes tend to be subclonal, implying subtype divergence at later stages of tumor evolution. Subclonality of recurrent HGSOC alterations is particularly evident for proliferative tumors, characterized by extreme genomic instability, absence of immune infiltration, and greater patient age. In contrast, differentiated tumors are characterized by largely intact genome integrity, high immune infiltration, and younger patient age. We propose an alternative model to discrete subtypes of HGSOC, in which tumors develop from an early differentiated spectrum to a late proliferative spectrum, along a timeline characterized by increasing genomic instability and subclonal expansion. The proposed methods provide a new approach to investigating tumor evolution through multi-omic analysis. Statement of SignificanceThis study proposes a method to infer whether transcriptome-based groupings of tumors differentiate early in carcinogenesis and are therefore potentially appropriate targets for therapy, and demonstrates that this is not the case for high-grade serous ovarian carcinoma (HGSOC). Significant findings for HGSOC include: O_LITumor purity, ploidy, and subclonality can be reliably inferred from different genomic platforms and show marked differences between subtypes C_LIO_LIRecurrent DNA alterations are associated with subtypes and tend to occur more frequently in subclones C_LIO_LISingle-cell sequencing of 42,000 tumor cells reveals widespread heterogeneity in tumor cell type composition that drives bulk subtype calls, but demonstrates a lack of intrinsic subtypes among tumor epithelial cells C_LIO_LIFindings prompt the dismissal of discrete transcriptome subtypes for HGSOC and replacement by a more realistic model of continuous tumor development that includes mixtures of subclones, accumulation of somatic aberrations, infiltration of immune and stromal cells in proportions correlated with tissue of origin and tumor stage, and evolution between properties previously associated with discrete subtypes C_LI

cancer biology↗

bioRxiv2018

The impact of stroma on the discovery of molecular subtypes and prognostic gene signatures in serous ovarian cancer

PurposeRecent efforts to improve outcomes for high-grade serous ovarian cancer, a leading cause of cancer death in women, have focused on identifying molecular subtypes and prognostic gene signatures, but existing subtypes have poor cross-study robustness. We tested the contribution of cell admixture in published ovarian cancer molecular subtypes and prognostic gene signatures.\n\nExperimental DesignPublic gene expression data, two molecular subtype classifications, and 61 published gene signatures of ovarian cancer were examined. Using microdissected data, we developed gene signatures of ovarian tumor and stroma. Computational simulations of increasing stromal cell proportion were performed by mixing gene expression profiles of paired microdissected ovarian tumor and stroma.\n\nResultsEstablished ovarian cancer molecular subtypes are strongly associated with the cell admixture. Tumors were classified as different molecular subtypes in simulations, when the percentage of stromal cells increased. Stromal gene expression in bulk tumor was weakly prognostic, and in one dataset, increased stroma was associated with anatomic sampling location. Five published prognostic gene signatures were no longer prognostic in a multivariate model that adjusted for stromal content alone.\n\nConclusionsThe discovery that molecular subtypes of high grade serous ovarian cancer is influenced by cell admixture, and stromal cell gene expression is crucial for interpretation and reproduction of ovarian cancer molecular subtypes and gene signatures derived from bulk tissue. Single cell analysis may be required to refine the molecular subtypes of high grade serous ovarian cancer. Because stroma proportion was weakly prognostic, elucidating the role of the tumor microenvironments components will be important.\n\nTranslational relevanceOvarian cancer is a leading cause of cancer death in women in the United States. Although the tumor responds to standard therapy for the majority of patients, it frequently recurs and becomes drug-resistant. Recent efforts have focused on identifying molecular subtypes and prognostic gene signatures of ovarian cancer in order to tailor therapy and improve outcomes. This study demonstrates that molecular subtype identification depends on the ratio of tumor to stroma within the specimen. We show that the specific anatomic location of the biopsy may influence the proportion of stromal involvement and potentially the resulting gene expression pattern. It will be crucial for these factors to be taken into consideration when interpreting and reproducing ovarian cancer molecular subtypes and gene signatures derived using bulk tissue and single cells. Furthermore, it will be important to define the relative proportions of stromal cells and model their prognostic importance in the tumor microenvironment.

bioinformatics↗

bioRxiv2018

The impact of different sources of heterogeneity on loss of accuracy from genomic prediction models

Cross-study validation (CSV) of prediction models is an alternative to traditional cross-validation (CV) in domains where multiple comparable datasets are available. Although many studies have noted potential sources of heterogeneity in genomic studies, to our knowledge none have system atically investigated their intertwined impacts on prediction accuracy across studies. We employ a hybrid parametric/non-parametric bootstrap method to realistically simulate publicly available compendia of microarray, RNA-seq, and whole metagenome shotgun (WMS) microbiome studies of health outcomes. Three types of heterogeneity between studies are manipulated and studied: imbalances in the prevalence of clinical and pathological covariates, 2) differences in gene covariance that could be caused by batch, platform, or tumor purity effects, and 3) differences in the \"true\" model that associates gene expression and clinical factors to outcome. We assess model accuracy while altering these factors. Lower accuracy is seen in CSV than in CV. Surprisingly, heterogeneity in known clinical covariates and differences in gene covariance structure have very limited contributions in the loss of accuracy when validating in new studies. However, forcing identical generative models greatly reduces the within/across study difference. These results, observed consistently for multiple disease outcomes and omics platforms, suggest that the most easily identifiable sources of study heterogeneity are not necessarily the primary ones that undermine the ability to accurately replicate the accuracy of omics prediction models in new studies. Unidentified heterogeneity, such as could arise from unmeasured confounding, may be more important.

bioinformatics↗

bioRxiv2017

Consensus on Molecular Subtypes of Ovarian Cancer

INTRODUCTIONVarious computational methods for gene expression-based subtyping of high-grade serous (HGS) ovarian cancer have been proposed. This resulted in the identification of molecular subtypes that are based on different datasets and were differentially validated, making it difficult to achieve consensus on which definitions to use in follow-up studies. We assess three major subtype classifiers for their robustness and association to outcome by a meta-analysis of publicly available expression data, and provide a classifier that represents their consensus.\n\nMETHODSWe use a compendium of 15 microarray datasets consisting of 1,774 HGS ovarian tumors to assess 1) concordance between published subtyping algorithms, 2) robustness of those algorithms to re-clustering across datasets, and 3) association of subtypes with overall survival. A consensus classifier is trained on concordantly classified samples, and validated by leave-one-dataset-out validation.\n\nRESULTSEach subtyping classifier identified subsets significantly differing in overall survival, but were not robust to re-fitting in independent datasets and grouped only approximately one third of patients concordantly into four subtypes. We propose a consensus classifier to identify the minority of unambiguously classifiable tumors across multiple gene expression platforms, using a 100-gene signature. The resulting consensus subtypes correlate with patient age, survival, tumor purity, and lymphocyte infiltration.\n\nCONCLUSIONSOur analysis demonstrates that most HGS ovarian cancers are not able to be subtyped. A minority of tumors can be classified and our proposed consensus classifier consolidates and improves on the robustness of three previously proposed subtype classifiers. It provides reliable stratification of patients with HGS ovarian tumors of clearly defined subtype, and will assist in studying the role of polyclonality in the majority of tumors that are not robustly classifiable.

cancer biology↗

bioRxiv2017

Transcriptome Deconvolution of Heterogeneous Tumor Samples with Immune Infiltration

Transcriptomic deconvolution in cancer and other heterogeneous tissues remains challenging. Available methods lack the ability to estimate both component-specific proportions and expression profiles for individual samples. We present DeMixT, a new tool to deconvolve high dimensional data from mixtures of more than two components. DeMixT implements an iterated conditional mode algorithm and a novel gene-set-based component merging approach to improve accuracy. In a series of experimental validation studies and application to TCGA data, DeMixT showed high accuracy. Improved deconvolution is an important step towards linking tumor transcriptomic data with clinical outcomes. An R package, scripts and data are available: https://github.com/wwylab/DeMixT/.

bioinformatics↗