Model Repository Research Articles

A generic detailed multibody model for simulating thoracic spine and ribcage kinematics

AbstractA reliable and comprehensive multibody musculoskeletal model of the thoracic spine and ribcage can offer valuable insight into the biomechanics of healthy and curved spines. In this study, we developed a generic rigid-body thoracic spine and ribcage model, which is kinematically determinate and controlled by spinal posture. A newly devised averaging constraint was implemented to model the kinematics of individual ribs and the sternum. Rib motion validation confirmed that the movement of adjacent ribs matched measured data across different tasks. We simulated 39 adolescent idiopathic scoliosis subjects aged 7–17 years, with an average (SD) Cobb angle of 15.9 (8.2) degrees. The average errors of multiple scoliosis metrics were less than 4 degrees for angle parameters, 5 mm for displacement parameters, and 3 percent for the ratio parameter. The model is straightforward and can simulate various daily activities (e.g., spinal articulation and breathing) while accurately capturing pathological deformities conforming to experimental data. It is available to the public on the open science platform Zenodo and will also be accessible through the AnyBody Managed Model Repository.

AI-PEDURO – Artificial intelligence in pediatric urology: Protocol for a living scoping review and online repository

BackgroundArtificial intelligence (AI) and machine learning (ML) methods are increasingly being applied in pediatric urology across a growing number of settings, with more extensive databases and wider interest for use in clinical practice. More than 30 ML models have been published in the pediatric urology literature, but many lack items required by contemporary reporting frameworks to be high quality. For example, most studies lack multi-institution validation, validation over time, and validation within the clinical environment, resulting in a large discrepancy between the number of models developed versus the number of models deployed in a clinical setting, a phenomenon known as the AI chasm. Furthermore, pediatric urology is a unique subspecialty of urology with low frequency conditions and complex phenotypes where clinical management can rely on a lower quality of evidence. ObjectiveTo establish the AI in PEDiatric UROlogy (AI-PEDURO) collaborative, which will carry out a living scoping review and create an online repository (www.aipeduro.com) for models in the field and facilitate an evidence synthesis of AI models in pediatric urology. Methods and AnalysisThe scoping review will follow PRISMA-ScR guidelines. We will include ML models identified through standardized search methods of four databases, hand-search papers, and user-submitted models. Retrieved records will be included if they involve ML algorithms for prediction, classification, or risk stratification for pediatric urology conditions. The results will be tabulated and assessed for trends within the literature. Based on data availability, models will be divided into clinical disease sections (e.g. hydronephrosis, hypospadias, vesicoureteral reflux). A risk assessment will be performed using the APPRAISE-AI tool. The retrieved model cards (brief summary model characteristics in table form) will be uploaded to the online repository for open access to clinicians, patients, and data scientists, and will be linked to the Digital Object Identifier (DOI) for each article. Discussion and ConclusionWe hope this living scoping review and online repository will offer a valuable reference for pediatric urologists to assess disease-specific ML models' scope, validity, and credibility to encourage opportunities for collaboration, external validation, clinical testing, and responsible deployment. In addition, the repository may aid in identifying areas in need of further research.

An analysis of FRE @ BC8 SympTEMIST track: named entity recognition.

This paper is a more in-depth analysis of the approaches used in our submission (Martínez A, García-Santa N. (2023) FRE @ BC8 SympTEMIST track: Named Entity Recognition Zenodo.) to the 'SympTEMIST' Named Entity Recognition (NER) shared subtask at 'BioCreative 2023'. We participated on the challenge submitting two systems based on a RoBERTa architecture LLM trained on Spanish-language clinical data available at 'HuggingFace' model repository. Before choosing the systems that would be submitted, we tried different combinations of the techniques described here: Conditional Random Fields and Byte-Pair Encoding dropout. In the second system we also included Sub-Subword feature based embeddings (SSW). The test set used in the challenge has now been released (López SL, Sánchez LG, Farré E etal. (2024) SympTEMIST Corpus: Gold Standard annotations for clinical symptoms, signs and findings information extraction. Zenodo), allowing us to analyze more in depth our methods, as well as measuring the impact of introducing data from CARMEN-I (Lima-López S, Farré-Maduell E, Krallinger M. (2023) CARMEN-I: Clinical Entities Annotation Guidelines in Spanish. Zenodo) corpus. Our experiments show the moderate effect of using the Sub-Subword feature based embeddings and the impact of including the symptom NER data from the CARMEN-I dataset. Database URL: https://physionet.org/content/carmen-i/1.0/.

Development and application of a population pharmacokinetic model repository for caffeine dose tailoring in preterm infants

ABSTRACT Background Considerable interindividual variability for the pharmacokinetics of caffeine in preterm infants has been demonstrated, emphasizing the importance of personalized dosing. This study aimed to develop and apply a repository of currently published population pharmacokinetic (PopPK) models of caffeine in preterm infants to facilitate model-informed precision dosing (MIPD). Research design and methods Literature search was conducted using PubMed, Embase, Scopus, and Web of Science databases. Relevant publications were screened, and their quality was assessed. PopPK models were reestablished to develop the model repository. Covariate effects were evaluated and the concentration–time profiles were simulated. An online simulation and calculation tool was developed as an instance. Results Twelve PopPK models were finally included in the repository. Preterm infants’ age and body size, especially the postnatal age and current weight, were identified as the most clinically critical covariates. Simulated blood concentration–time profiles across these models were comparable. Caffeine citrate-dose regimen should be adjusted according to the age and body size of preterm infants. The developed online tool can be used to facilitate clinical decision-making. Conclusions The first developed repository of PopPK models for caffeine in preterm infants has a wide range of potential applications in the MIPD of caffeine.

"Tear down that wall"-a critical evaluation of bioinformatic resources available for lysin researchers.

Lytic enzymes, or lysins for short, break down peptidoglycan and interrupt the continuity of the cell wall, which, in turn, causes osmotic lysis of the bacterium. Their ability to destroy bacteria from within makes them promising antimicrobial agents that can be used as alternatives or supplements to antibiotics. In this paper, we briefly summarize basic terms and concepts used to describe lysin sequences and delineate major lysin groups. More importantly, we describe the domain repertoire found in lysins and critically review bioinformatic tools or databases which are used in studies of these enzymes (with particular emphasis on the repositories of Hidden Markov models). Finally, we present a novel comprehensive, meticulously curated set of lysin-related family and domain models, sort them into clusters that reflect major families, and demonstrate that the selected models can be used to efficiently search for new lysins.

Automatic generation of system model diagrams driven by multi-source heterogeneous data

The demand for functionality, performance, safety, and reliability in complex product domains is continuously increasing, leading to growing product complexity. Traditional document-based systems engineering approaches face numerous challenges in terms of communication efficiency, traceability, and maintenance when dealing with complex product system design. Consequently, Model-Based Systems Engineering (MBSE) has been widely adopted. To improve the efficiency of system modelling, an effective method is needed to automate the process by reusing existing design resources. This study proposes a knowledge graph-based approach for the automatic generation of system model diagrams, consisting of three main steps. Firstly, a system model graph is constructed using an existing system model repository. Then, the required modelling elements are extracted from design documents, and reusable model elements are obtained based on the system model graph. Lastly, a mapping relationship is established between system model elements extracted from multi-source heterogeneous data and SysML metamodel, enabling the automatic generation of system model diagrams based on SysML views. This approach effectively leverages the information from the existing system model repository and achieves automated mapping conversion from multi-source heterogeneous data to the SysML system model. A communication satellite case study is presented to demonstrate the capability of this method.

UVLHub: A feature model data repository using UVL and open science principles

Feature models are the de facto standard for modelling variabilities and commonalities in features and relationships in software product lines. They are the base artefacts in many engineering activities, such as product configuration, derivation, or testing. Concrete models in different domains exist; however, many are in private or sparse repositories or belong to discontinued projects. The dispersion of knowledge of feature models hinders the study and reuse of these artefacts in different studies. The Universal Variability Language (UVL) is a community effort textual feature model language that promotes a common way of serializing feature models independently of concrete tools. Open science principles promote transparency, accessibility, and collaboration in scientific research. Although some attempts exist to promote feature model sharing, the existing solutions lack open science principles by design. In addition, existing and public feature models are described using formats not always supported by current tools. This paper presents ▪ , a repository of feature models in UVL format. ▪ provides a front end that facilitates the search, upload, storage, and management of feature model datasets, improving the capabilities of discontinued proposals. Furthermore, the tool communicates with Zenodo – one of the most well-known open science repositories – providing a permanent save of datasets and following open science principles. ▪ includes existing datasets and is readily available to include new data and functionalities in the future. It is maintained by three active universities in variability modelling.

The AI community building the future? A quantitative analysis of development activity on Hugging Face Hub

Open model developers have emerged as key actors in the political economy of artificial intelligence (AI), but we still have a limited understanding of collaborative practices in the open AI ecosystem. This paper responds to this gap with a three-part quantitative analysis of development activity on the Hugging Face (HF) Hub, a popular platform for building, sharing, and demonstrating models. First, various types of activity across 348,181 model, 65,761 dataset, and 156,642 space repositories exhibit right-skewed distributions. Activity is extremely imbalanced between repositories; for example, over 70% of models have 0 downloads, while 1% account for 99% of downloads. Furthermore, licenses matter: there are statistically significant differences in collaboration patterns in model repositories with permissive, restrictive, and no licenses. Second, we analyse a snapshot of the social network structure of collaboration in model repositories, finding that the community has a core-periphery structure, with a core of prolific developers and a majority of isolate developers (89%). Upon removing these isolates from the network, collaboration is characterised by high reciprocity regardless of developers’ network positions. Third, we examine model adoption through the lens of model usage in spaces, finding that a minority of models, developed by a handful of companies, are widely used on the HF Hub. Overall, the findings show that various types of activity across the HF Hub are characterised by Pareto distributions, congruent with open source software development patterns on platforms like GitHub. We conclude with recommendations for researchers, and practitioners to advance our understanding of open AI development.

MakeSBML: a tool for converting between Antimony and SBML.

We describe a web-based tool, MakeSBML (https://sys-bio.github.io/makesbml/), that provides an installation-free application for creating, editing, and searching the Biomodels repository for SBML-based models. MakeSBML is a client-based web application that translates models expressed in human-readable Antimony to the System Biology Markup Language (SBML) and vice-versa. Since MakeSBML is a web-based application it requires no installation on the user's part. Currently, MakeSBML is hosted on a GitHub page where the client-based design makes it trivial to move to other hosts. This model for software deployment also reduces maintenance costs since an active server is not required. The SBML modeling language is often used in systems biology research to describe complex biochemical networks and makes reproducing models much easier. However, SBML is designed to be computer-readable, not human-readable. We therefore employ the human-readable Antimony language to make it easy to create and edit SBML models.

Patient-Derived Conditionally Reprogrammed Cells in Prostate Cancer Research.

Prostate cancer (PCa) remains a leading cause of mortality among American men, with metastatic and recurrent disease posing significant therapeutic challenges due to a limited comprehension of the underlying biological processes governing disease initiation, dormancy, and progression. The conventional use of PCa cell lines has proven inadequate in elucidating the intricate molecular mechanisms driving PCa carcinogenesis, hindering the development of effective treatments. To address this gap, patient-derived primary cell cultures have been developed and play a pivotal role in unraveling the pathophysiological intricacies unique to PCa in each individual, offering valuable insights for translational research. This review explores the applications of the conditional reprogramming (CR) cell culture approach, showcasing its capability to rapidly and effectively cultivate patient-derived normal and tumor cells. The CR strategy facilitates the acquisition of stem cell properties by primary cells, precisely recapitulating the human pathophysiology of PCa. This nuanced understanding enables the identification of novel therapeutics. Specifically, our discussion encompasses the utility of CR cells in elucidating PCa initiation and progression, unraveling the molecular pathogenesis of metastatic PCa, addressing health disparities, and advancing personalized medicine. Coupled with the tumor organoid approach and patient-derived xenografts (PDXs), CR cells present a promising avenue for comprehending cancer biology, exploring new treatment modalities, and advancing precision medicine in the context of PCa. These approaches have been used for two NCI initiatives (PDMR: patient-derived model repositories; HCMI: human cancer models initiatives).

Repository of MRI-derived models of the breast with single and multiple benign and malignant tumors for microwave imaging research.

The diagnosis of breast cancer through MicroWave Imaging (MWI) technology has been extensively researched over the past few decades. However, continuous improvements to systems are needed to achieve clinical viability. To this end, the numerical models employed in simulation studies need to be diversified, anatomically accurate, and also representative of the cases in clinical settings. Hence, we have created the first open-access repository of 3D anatomically accurate numerical models of the breast, derived from 3.0T Magnetic Resonance Images (MRI) of benign breast disease and breast cancer patients. The models include normal breast tissues (fat, fibroglandular, skin, and muscle tissues), and benign and cancerous breast tumors. The repository contains easily reconfigurable models which can be tumor-free or contain single or multiple tumors, allowing complex and realistic test scenarios needed for feasibility and performance assessment of MWI devices prior to experimental and clinical testing. It also includes an executable file which enables researchers to generate models incorporating the dielectric properties of breast tissues at a chosen frequency ranging from 3 to 10 GHz, thereby ensuring compatibility with a wide spectrum of research requirements and stages of development for any breast MWI prototype system. Currently, our dataset comprises MRI scans of 55 patients, but new exams will be continuously added.

Multi-objective model transformation chain exploration with MOMoT

Context:The increasing complexity of modern systems leads to an increasing amount of artifacts that are used along the model-based software and systems development lifecycle. This also includes model transformations, which serve for mapping models between representations, e.g., for verification and validation purposes. Objectives:Model repositories manage this variety of artifacts and promote reusability, but should also enable the bundling of compatible artifacts. Therefore, model transformations should be reused and arranged into transformation chains to support more complex transformation scenarios. The resulting transformation should correspond to the user’s interest in terms of quality criteria such as model coverage, transformation coverage, and number of transformation steps, thus assembling such chains becomes a multi-objective problem. Methods:A novel multi-objective approach for exploring possible transformation chains residing in model repositories is presented. MOMoT, a model-driven optimization framework, is leveraged to explore the transformation space spanned by the repository. For demonstration, three differently populated repositories are considered. Results:We have extended MOMoT with an exhaustive, multi-objective search that explores the entire model transformation space defined by graph transformation rules, allowing all possible transformation chains to be considered as solution. Accordingly, the optimal solutions were identified in the demonstration cases with negligible computation time. Conclusion:The approach assists modelers when there are multiple chains for transforming an input model to a specified output model to consider. Our evaluation shows that the approach elicits all legitimate transformation chains, thus enabling the modelers to consider trade-offs in view of multiple criteria selection.

Advanced discovery mechanisms in model repositories

SummaryAs model‐driven engineering gains traction and poses as the new paradigm for software engineering, it raises a need for efficient approaches and tools to manage, discover, and retrieve relevant modeling artifacts. Hence, industry and academia are conceiving effective ways to store, search, and retrieve heterogeneous model artifacts that employ advanced discovery mechanisms. This paper presents MDEForge‐Search, a novel approach to discovering heterogeneous model artifacts over MDEForge, a distributed cloud‐based model repository. We designed advanced discovery mechanisms that retrieve heterogeneous artifacts within their context (megamodel) and reuse them across model management services. In addition, a domain‐specific approach has been proposed to formulate queries in terms of keywords, search tags, conditional operators, quality model assessment services and a transformation chain discoverer. Finally, the applicability of our approach was assessed in a recommender system modeling framework, which, thanks to the operated integration, can rely on the availability of more than 5000 model artifacts currently persisted in our cloud‐based model repository.

The simulation experiment description markup language (SED-ML): language specification for level 1 version 5.

Modern biological research is increasingly informed by computational simulation experiments, which necessitate the development of methods for annotating, archiving, sharing, and reproducing the conducted experiments. These simulations increasingly require extensive collaboration among modelers, experimentalists, and engineers. The Minimum Information About a Simulation Experiment (MIASE) guidelines outline the information needed to share simulation experiments. SED-ML is a computer-readable format for the information outlined by MIASE, created as a community project and supported by many investigators and software tools. Level 1 Version 5 of SED-ML expands the ability of modelers to define simulations in SED-ML using the Kinetic Simulation Algorithm Onotoloy (KiSAO). While it was possible in Version 4 to define a simulation entirely using KiSAO, Version 5 now allows users to define tasks, model changes, ranges, and outputs using the ontology as well. SED-ML is supported by a growing ecosystem of investigators, model languages, and software tools, including various languages for constraint-based, kinetic, qualitative, rule-based, and spatial models, and many simulation tools, visual editors, model repositories, and validators. Additional information about SED-ML is available at https://sed-ml.org/.

A probabilistic neural twin for treatment planning in peripheral pulmonary artery stenosis.

The substantial computational cost of high-fidelity models in numerical hemodynamics has, so far, relegated their use mainly to offline treatment planning. New breakthroughs in data-driven architectures and optimization techniques for fast surrogate modeling provide an exciting opportunity to overcome these limitations, enabling the use of such technology for time-critical decisions. We discuss an application to the repair of multiple stenosis in peripheral pulmonary artery disease through either transcatheter pulmonary artery rehabilitation or surgery, where it is of interest to achieve desired pressures and flows at specific locations in the pulmonary artery tree, while minimizing the risk for the patient. Since different degrees of success can be achieved in practice during treatment, we formulate the problem in probability, and solve it through a sample-based approach. We propose a new offline-online pipeline for probabilistic real-time treatment planning which combines offline assimilation of boundary conditions, model reduction, and training dataset generation with online estimation of marginal probabilities, possibly conditioned on the degree of augmentation observed in already repaired lesions. Moreover, we propose a new approach for the parametrization of arbitrarily shaped vascular repairs through iterative corrections of a zero-dimensional approximant. We demonstrate this pipeline for a diseased model of the pulmonary artery tree available through the Vascular Model Repository.

Abstract 6915: Integrated proteogenomic analysis for the NCI patient-derived cancer model repository

Abstract Background: Studying protein expression involved in cancer from patient-derived xenografts (PDX) enhances understanding of tumorigenesis and may shed light on patient therapeutic decisions. Recently, mass spectrometry-based proteomics sequencing techniques have enabled the acquisition of high-throughput PDX protein/phosphoprotein expression profiles and paved a way for investigating oncogenic signaling pathways and their interactions in different tumor histologies. Integrated analysis of these proteomic data with existing biomarkers derived from whole exome and RNAseq data will provide the community with a rich resource for translational research. Materials and methods: iTRAQ-based proteomics/phospho-proteomics data, with 11 Tandem Mass Tag (TMT) channels, were processed and median-normalized at gene level. The gene expression data were derived from tximport and DESeq2 from of RNA-Seq. In addition, whole exome sequencing (WES) data were used to compute mutations and copy number profiles. For data visualization, the R Bioconductor package ComplexHeatmap was used extensively. Results: PDX proteomics/phospho-proteomics data were analyzed alongside with gene expression, copy number and mutational profiles for several cancer types, revealing vivid molecular regulation for many cancer-related pathways. A moderate yet significant Spearman correlation coefficient of over 0.45 and 0.5 was observed between proteomics and RNA-Seq data at gene and pathway level, respectively. The PDX protein and gene expression pattern from common cancer regulatory pathways of bladder (BLCA, n=32 models), colon (COAD, n=82), lung (NSCLC, n=32), pancreatic (PAAD, n=52), metastatic breast cancer (CSNOS) (n=16) and other histologies were aggregated and discussed at the model level. EGFR, MET and cell-cycle genes CDKN2A and CDKN2B exhibited significant positive correlation among copy number, gene expression and protein expression profiles. Protein and gene expression profiles of PAM50 genes were shown for CSNOS samples, to further sub-classify or correct misclassification. Further analysis using proteomics/phospho-proteomics data with CNA identified several trans-regulatory events. Lastly, we identified RTN1 as a potential biomarker, exhibiting lower expression in sensitive rare-tumor PDX models prior to Axitinib and Vandetanib treatment. Conclusions: With the multi-omics datasets comprising proteomics/phospho-proteomics, RNA-Seq and WES datasets from the NCI Patient Derived Models Repository cohort, we were able to query some important cancer biological processes at a higher resolution. In addition, we revealed gene- and pathway-level regulatory differences from various histologies. Overall, the multi-omics data from PDX models showed promising recapitulation of original tumor activity and should continue to serve as an amenable and scalable drug screening platform for pre-clinical trials. Citation Format: Peter I. Wu, Li Chen, Yuri Kotliarov, Jianwen Fang, Yingdong Zhao, Yvonne Evrard, Lijun Chen, Shahanawaz Jiwani, Biswajit Das, Chris A. Karlovich, Hui Zhang, Lisa McShane, Melinda G. Hollingshead, Mickey Williams, James H. Doroshow. Integrated proteogenomic analysis for the NCI patient-derived cancer model repository [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6915.

Abstract 6898: Cholangiocarcinoma patient-derived models in the NCI Patient-Derived Models Repository

Abstract Cholangiocarcinoma (CHOL) is an aggressive rare malignancy arising in the biliary tract with a 5-year relative survival rate of 10% and a 60-70% recurrence rate following surgical intervention. Advances in treatment strategies are hampered by the lack of well annotated and characterized preclinical models. The National Cancer Institute’s Patient-Derived Models Repository (NCI PDMR; https://pdmr.cancer.gov) has developed a national repository of Patient-Derived Models (PDMs) comprised of patient-derived xenografts (PDXs), in vitro patient-derived tumor cell cultures (PDCs) and cancer associated fibroblasts (CAFs) as well as patient-derived organoids (PDOrg). These PDMs are clinically annotated with molecular information available in an easily accessible database for the extramural community. To date, 94 patient CHOL tumor specimens (resections, biopsies, rapid autopsy) have been received from 55 unique patients with an overall PDX take rate of 36.3% (80 assessable specimens). There are currently 15 CHOL PDX models available for the research community to request with 14 more in final QC. Nine of the 15 public PDX models were generated from unique lesions in two rapid autopsy patients and while histo-morphologically consistent, some genomic heterogeneity is observed between the unique anatomical locations likely due to tumor evolution during metastasis. In addition, 27 in vitro PDOrgs, PDCs, and CAFs have been generated from patient or PDX material matched to most of the existing PDX models to allow for comparative translation research. The genetic landscape of the CHOL PDXs includes oncogenic drivers in NRAS, IDH1, ERBB2, TP53, KRAS and FGFR2 fusions; OncoKB likely oncogenic variants including ARID1A, AXIN1 and BAP1; and CDKN2A deep deletions representing most of the common alterations in CHOL. Early passage preclinical PDX and in vitro models for CHOL, and many other cancer types, with translational relevant features along with associated NextGen sequencing and patient treatment history are available from the NCI PDMR. Availability of preclinical models that can be used to study these cancers and improve preclinical drug screening is of high importance to better understand the biology of these cancers and prioritize novel therapeutics from bench to clinic. Funded by NCI Contract No. HHSN261200800001E Citation Format: Yvonne A. Evrard, Cindy R. Timme, Biswajit Das, Gareth Bliss, Carrie Bonomi, Carley Border, Ting-Chia Chang, Alice Chen, Li Chen, Michelle A. Crespo-Eugeni, Kevin Cooper, Natalie Czarra, Isabella Czernia, Kelly Dougherty, Marion Gibson, Tara Grinnage-Pulley, Shahanawaz Jiwani, Keegan Kalmbach, Chris A. Karlovich, Chelsea McGlynn, Michael Mullendore, Matthew Murphy, Rini Pauly, Kevin Plater, Jessica Steed, Luke Stockwin, Shannon Uzelac, Peter I-Fan Wu, Dianne L. Newton, P. Mickey Williams, Melinda G. Hollingshead, James H. Doroshow. Cholangiocarcinoma patient-derived models in the NCI Patient-Derived Models Repository [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6898.

Abstract 6911: Expanding reported molecular characteristics for the NCI Patient-Derived Models Repository

Abstract The National Cancer Institute’s Patient-Derived Models Repository (NCI PDMR; https://pdmr.cancer.gov) has developed a national repository of Patient-Derived Models (PDMs) comprised of patient-derived xenografts (PDXs), in vitro patient-derived tumor cell cultures (PDCs) and cancer associated fibroblasts (CAFs) as well as patient-derived organoids (PDOrg). These PDMs are clinically annotated with molecular information available in an easily accessible database for the extramural community. To ensure the data and models remains pertinent to current translational efforts, NCI actively reviews and adds new molecular features to the public database for researchers to use. While these data could be obtained through analysis of the available NextGen sequencing files in the PDMR database, many laboratories do not have the bioinformatics support to run these assessments for models that are of research interest. Therefore, new pipelines for an array of additional model and molecular characteristics present in PDX, PDOrg, and PDC models have been validated and are reported in the public database in a searchable format. Current model characteristics include addition of OncoTree Codes as a second histologic classifier, use of the PDXNet Consortium’s inferred ancestry caller to better represent Native/Latin American populations, human leukocyte antigen (HLA) typing, molecular classification of tumors (e.g. CMS subtyping of colorectal adenocarcinoma, PAM50 signature based subtyping of breast carcinoma), microsatellite instability (MSI) assessment, and identification of clinically relevant fusions. SOPs for the pipelines used are available on the public website to allow for independent validation and use by other laboratories. Funded by NCI Contract No. HHSN261200800001E Citation Format: Cindy R. Timme, Yvonne A. Evrard, Biswajit Das, Ting-Chia Chang, Alice Chen, Li Chen, Michelle A. Crespo-Eugeni, Shahanawaz Jiwani, Chris A. Karlovich, Rini Pauly, Peter I-Fan Wu, Dianne L. Newton, P. Mickey Williams, Melinda G. Hollingshead, James H. Doroshow. Expanding reported molecular characteristics for the NCI Patient-Derived Models Repository [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6911.

Abstract 227: Novel and easy-to-use approach to triple-negative breast cancer tumoroid culture

Abstract Triple-negative breast cancer (TNBC) accounts for 15-20% of all breast cancer diagnoses in the United States. Patients with metastatic TNBC have a low survival rate, driven in part by limited treatment options for the disease. Tumoroids, also known as cancer organoids, have been shown to better maintain tumor-specific mutational and gene expression profiles over long-term culture, and evidence is growing that the functional response of tumoroids to therapeutics better recapitulates in vivo response. These improved preclinical models of TNBC may facilitate the development of new therapies, but tumoroids are yet to be widely adopted due to challenging culture methods. In this study, we show that two TNBC tumoroid models publicly available through the National Cancer Institute Patient-Derived Models Repository can be expanded in suspension culture using Gibco™ OncoPro™ Tumoroid Culture Medium supplemented with FGF10 and beta-estradiol. We compared the bulk tumoroid morphologies, growth rates, maintenance of mutational and gene expression profiles, and molecular phenotypes of cells grown in the NCI-recommended medium in embedded culture, and the OncoPro medium in embedded culture and suspension culture. We saw comparable morphologies and growth rates in all culture conditions. A multivariate analysis of the allelic frequencies of single nucleotide variants and ploidy values of copy number variants for cells expanded in all experimental conditions revealed >95% and >88% correlation, respectively, with the initial material. Interestingly, despite the change in the medium formulation, gene expression was also highly correlated between the initial material and the OncoPro medium conditions for >20,000 RefSeq genes (Pearson r>0.94) and for a cancer-gene specific panel of 1,423 genes (r>0.87). Further analysis showed unchanged expression levels in 97% of total genes analyzed and the basal molecular subtype was maintained for both lines in all conditions. The OncoPro media system was also used to derive a new TNBC tumoroid line, which demonstrates consistent growth rate and maintains the patient-specific mutational profile of the original breast tumor. Differential gene expression analysis revealed that changes in the gene expression profile were largely due to the loss of non-malignant cell populations from the original tumor. Finally, we utilized this new tumoroid line to examine donor-specific killing efficiencies of primary natural killer cells in a co-culture experiment. The discrepancy in the dynamics of killing a tumoroid line versus the standard K562 cell line indicates that tumoroids may offer an improved approach for screening new cell therapies. In summary, the easy-to-adopt OncoPro medium and culture method does not change the characteristics of existing models and may be used to derive new tumoroid lines, both of which can be leveraged for preclinical studies to develop critical new therapies for TNBC. Citation Format: Brittany Balhouse, Chris Yankaskas, Colin Paul, Shyanne Salen, Sylvia Beam, Pradip Shahi Thakuri, Matt Dallas, David Kuninger. Novel and easy-to-use approach to triple-negative breast cancer tumoroid culture [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 227.

Abstract 6768: Matched XPDX™derivative models as a platform for continuity across preclinical drug development

Abstract Drug discovery is a reductive process in which druglike entities are carried through increasingly complex assay formats to tease out the compounds best suited for human studies. In vivo efficacy remains the benchmark for successful preclinical drug development. Cell-based models are a mainstay in early stages but translatability of different models across platforms remains a challenge. In oncology, patient-derived xenografts (PDX) are the gold standard for animal studies because they recapitulate key features of human tumors including the cellular heterogeneity, growth patterns, and drug response of the original patient tumor. Ex vivo, PDX derivatives have the potential to harness these same traits thereby extending the utility of PDX models to earlier stages of preclinical drug development. The KIYA-Predict™ platform is a clinically validated tool for the assessment of ex vivo drug responses across a broad range of drug classes that when applied to XPDX models shows good correlation to in vivo drug responses and excellent intra-sample reproducibility over time. Working with the well-characterized XenoSTART™ XPDX model repository, Kiyatec has derived 2D cell lines and organotypic 3D-expansion cultures to complement the ex vivo predictive value of our KIYA-PREDICT™ XPDX drug response profiling (DRP) platform. The development of syngeneic models applicable to simpler, high-throughput library screening (2D formats) and lower-throughput more complex mechanistic studies (3D organotypic models) could alleviate the burden and pitfalls of extrapolating data across unrelated models. Matched sets of 2D cell lines (2D-XPDX™) and 3D organotypic (XPDXO™) cultures were established from single cell suspensions of digested, mouse-cell depleted XPDX™ tissue samples. Drug response profiling of the XPDX™ in their various ex vivo culture formats showed relative similarities across platforms. Genomic profiling of the derivates confirmed shared tissue origin and retention of genomic and phenotypic signatures to varying degrees across models. Mechanistic evaluation of 3D organotypic culture demonstrated that disease-relevant biomarkers and processes are maintained. Together these results demonstrate the feasibility of this novel drug discovery platform that avoids ambiguity imparted by discordant models. Citation Format: Melissa Millard, Aaron Carlson, Danielle Y. Nadeau, Kimberly J. Burgess, Alyssa Simonson, Armando Diaz, Michael Wick, Teresa DesRochers. Matched XPDX™derivative models as a platform for continuity across preclinical drug development [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6768.