Science/AAAS


Session Type Key

In-Person Sessions
Hybrid Sessions
Virtual Sessions

$$ - denotes additional registration fee required to attend session

Reset

Advanced Filters

Computational Chemogenomics: Bridging Chemical and Biological Spaces:
08:00am - 11:45am USA / Canada - Central - August 22, 2022 | Location: Henry Clarke (Marriott Marquis Chicago)
Dr. Jose L Medina-Franco, Organizer, Presider
Division: [CINF] Division of Chemical Information
Session Type: Oral - Hybrid
Division/Committee: [CINF] Division of Chemical Information

The symposium will present the progress on the development and application of computational methods to characterize, explore and expand the chemical space and biological spaces and their intersection. The methods and practical applications in computational chemogenomics include but are not limited to artificial intelligence, data mining and visualization, and library design. Applications of virtual screening and inverse virtual screening (target fishing) for drug repurposing and multi-target drug discovery are welcome.

Monday
Introductory Remarks
08:00am - 08:05am USA / Canada - Central - August 22, 2022 | Location: Henry Clarke (Marriott Marquis Chicago)
Division: [CINF] Division of Chemical Information
Session Type: Oral - Hybrid

Monday
3755661 - Biomedical knowledge graphs and their application to drug discovery
08:05am - 08:30am USA / Canada - Central - August 22, 2022 | Location: Henry Clarke (Marriott Marquis Chicago)
Division: [CINF] Division of Chemical Information
Session Type: Oral - Hybrid
The volume of biomedical research data stored in various databases has grown immensely in recent years due to the proliferation of high-throughput biomedical ‘-omics’ technologies. Nearly all of respective databases, or ‘knowledge sources’ (KSs), address a particular area of biomedical research, leading to natural diversity but also growing disintegration between individual KSs, which generates downstream inefficiencies when mining diverse databases for knowledge discovery. Expanding efforts, both in academia and industry, are focused on the development of methods and tools to enable semantic integration and concurrent exploration of disparate biomedical KSs, using specially constructed biomedical ‘graph knowledgebases’ (GKBs) that support the generation of new knowledge through the application of reasoning tools and algorithms. Our group has contributed to these efforts by initiating the development of a GKB-based question-answering system termed Reasoning Over Biomedical Objects linked in Knowledge-Oriented Pathways (ROBOKOP) 1,2. ROBOKOP’s publicly accessible user interface (UI) 3 allows users to address both relatively simple questions such as “what genes are associated with drug-induced liver injury?” and more complex ones such as “What biological mechanisms are responsible for drug action?” I will discuss the development of ROBOKOP and provide examples of applications including the elucidation of Adverse or Clinical Outcome Pathways of drug action, development of domain-specific KG such as COVID-KOP 4, and drug repurposing including methodologies relying on KG embeddings 5 and machine learning.
Monday
3748340 - Prediction of Xenobiotic localization at subcellular level.
08:30am - 08:55am USA / Canada - Central - August 22, 2022 | Location: Henry Clarke (Marriott Marquis Chicago)
Division: [CINF] Division of Chemical Information
Session Type: Oral - Hybrid
Understanding subcellular distribution and the mechanism of xenobiotics can help in modulating subcellular dysfunction mediated diseases. Therefore, with improved knowledge of how xenobiotics are distributed across subcellular locations and the mechanism for a specific molecule can play a crucial role in assessing drug efficacy and toxicity. Such knowledge would widen therapeutic windows by allowing specific receptors to be targeted efficiently. Based on datasets that provide information on the subcellular locations of proteins and their ligands, we developed machine learning models for 42 subcellular locations. Such models were trained and validated based on the grid search method and the best models based on Cohen’s Kappa scores were selected.
We have trained 22300+ machine learning models employing 19 different fingerprint-based features and 28 different ML classifiers for 42 different subcellular locations. All the data and models generated from the project are made available as open-source and user-friendly webportal.

Monday
3755175 - The dark cancer kinome: Untapped opportunities to target understudied kinases in cancer
08:55am - 09:20am USA / Canada - Central - August 22, 2022 | Location: Henry Clarke (Marriott Marquis Chicago)
Stephan Schuerer, Presenter
Division: [CINF] Division of Chemical Information
Session Type: Oral - Hybrid
The approval of the first kinase inhibitor, Gleevec, in 2001, ushered in a paradigm shift for oncological treatment—the use genomic data for targeted, efficacious therapies. Since then, about 60 small molecule kinase inhibitors have been approved, solidifying the case for kinases as a highly druggable and attractive target class. Despite the established role deregulated kinase activity plays in cancer, only a small subset of the Kinome has been effectively “drugged”. Moreover, a quarter of the more than 600 human kinases are vastly understudied, as formally defined in the Illuminating the Druggable Genome Project (IDG, https://druggablegenome.net).
We have developed a comprehensive scoring system, the Clinical Kinase Index (CKI, http://cki.ccs.miami.edu/), which utilizes differential gene expression, clinical and pathological parameters, overall survival and mutational hotspot analyses to rank and prioritize clinically-relevant kinase targets across 17 solid tumor cancers from The Cancer Genome Atlas. These findings suggest that dark kinases have potential clinical value as biomarkers or as new drug targets, which warrant further study.
To identify small molecules that directly target and inhibit dark kinases, for the development of chemical probes or drug leads, we have developed a pipeline that combines multi-task deep neural network models trained on activity data from across the Kinome with structure-based molecular simulations.
We have applied this computational workflow to identify novel small molecule for dark kinases with no known small molecule binders and no protein structure. For a novel pseudokinase with no data, we have now developed advanced lead compounds.

Monday
3752405 - Visualization and analysis of large collection of chemical libraries using Generated Topographic Mapping
09:20am - 09:45am USA / Canada - Central - August 22, 2022 | Location: Henry Clarke (Marriott Marquis Chicago)
Division: [CINF] Division of Chemical Information
Session Type: Oral - Hybrid
Here, we demonstrate how different chemical libraries can be compared using Generative Topographic Mapping (GTM)1. GTM is a dimensionality reduction technique, which allows to visualize on 2D map chemical structures representing by N molecular descriptors. Due to probabilistic nature of GTM, any chemical library can be encoded by cumulated responsibility (CR) vector identifying the data probability distribution in a studied chemical space according to chemical structure of molecules. Any physical property or biological activity of a given library can be accounted for in a GTM activity landscape2, which, in turn, can be encoded by activity weighted cumulated responsibility (awCR) vector, characterizing data probability distribution according to both structure and activity. The coverage of two different libraries can be assessed by comparison of their CR (or awCR) vectors. Ensemble of M libraries can be visualized on a meta-GTM using related CR (awCR) vectors as an input for the map construction instead of molecular descriptors. This approach has been applied to two library collections: (i) 37 chemical suppliers libraries containing about 2 M compounds and (ii) 2500 DNA-Encoded Libraries (DEL) containing 2.5B compounds3. We have demonstrated that similarity between two selected libraries depends not only on their chemical composition but varies as a function of considered activities.
Monday
Intermission
09:45am - 10:00am USA / Canada - Central - August 22, 2022 | Location: Henry Clarke (Marriott Marquis Chicago)
Division: [CINF] Division of Chemical Information
Session Type: Oral - Hybrid

Monday
Exogenous metal particles and ions from implant devices are known to cause severe toxic events with symptoms ranging from adverse local tissue reactions to systemic toxicities, potentially leading to the development of cancers, heart conditions, and neurological disorders. Toxicity mechanisms, also known as Adverse Outcome Pathways (AOPs), that explain these metal-induced toxicities are severely understudied. In addition, the cost and low throughput of experimental testing for metal toxicity presents substantial barriers for rapid regulatory assessment of novel metal-containing products. We employed a combination of modern knowledge mining and protein structure-based modeling approaches to develop a unique metallomics database and identify the proteome-level perturbations caused by chromium, cobalt, molybdenum, nickel, and titanium along with pathways that link these events to human diseases. We captured 177 protein-metal ion complex structures representing direct metal-protein interactions. We also identified 347 metal-protein connections through knowledge graph mining and used the graph to impute another 402 relationships. We used a database of secreted human proteins to prioritize 72 proteins hypothesized to directly contact implant surfaces and contribute to metal implant-related adverse outcomes. We have considered three scenarios of AOPs elucidation: (i) the metal-protein-disease relationship was previously known; (ii) the metal-protein, protein-disease, and metal-disease relationships were individually known but were not unified in an AOP; and (iii) one out of three relationships was unknown and was imputed by our methods. These scenarios were illustrated by three respective case studies on nickel-induced allergy/hypersensitivity, cobalt-induced congestive heart failure, and titanium-induced periprosthetic osteolysis. Our studies resulted in the first public metallomics database, a novel data-integration strategy based on creating a metal-centric biomedical knowledge graph, analytical workflows to explore metallomics data to elucidate metal AOPs, and, finally, testable hypotheses concerning biological mechanisms of metal toxicity. Studies reported herein represent an innovative, integrated computational approach to enable focused toxicity testing of metal-containing implants using non-animal methods. All workflows, data, and results produced in this study are publicly available at https://github.com/DnlRKorn/Knowledge_Based_Metallomics/.
Monday
3741523 - Discovery and characterization of high affinity peptide-based macrocyclic ligands with computational approaches
10:25am - 10:50am USA / Canada - Central - August 22, 2022 | Location: Henry Clarke (Marriott Marquis Chicago)
Lela Vukovic, Presenter
Division: [CINF] Division of Chemical Information
Session Type: Oral - Hybrid
Discovery of molecules that inhibit specific protein-protein interactions (PPI) is needed for developing new therapeutic interventions. Chemically modified peptides (peptidomimetics) are among the most promising candidates for PPI modulation. Combinatorial libraries of peptidomimetic compounds can be synthesized with molecular discovery display technologies. By panning the prepared libraries against target proteins, thousands of promising peptidomimetic ligands with high target affinities are usually identified by the affinity-selection process. However, narrowing down thousands of ligands and discovering those ligands with the highest affinities for target proteins remains a critical challenge. Here, we used docking calculations and the atomistic molecular dynamics simulations to examine sets of peptidomimetic ligands identified by the affinity-selection process for target proteins. Then, we developed new sequence and chemical space analysis tools of affinity-selected ligands from experimental datasets and applied it to example datasets. Lastly, we developed new machine learning models, trained on sequence and simulation-based features, to predict new ligands of the selected target proteins. Our simulations and analyses characterize factors that contribute to peptidomimetic ligand recognition of target proteins on the level of whole datasets of affinity-selected molecules. Furthermore, our machine learning codes and models, which can be applied to a wide range of protein targets, may be capable of reducing the time and cost to yield new peptide-based drugs.
Monday
3755642 - Navigating through chemical and biological space with extended similarity methods
10:50am - 11:15am USA / Canada - Central - August 22, 2022 | Location: Henry Clarke (Marriott Marquis Chicago)
Division: [CINF] Division of Chemical Information
Session Type: Oral - Hybrid
We will discus the foundation and some of the key applications of our recently introduced extended similarity indices, that allow us to compare an arbitrary number of molecules simultaneously. These new similarity indices have two key advantages: they allow us to quantify correlations between any number of molecules, and they do so with linear scaling. We will see how these indices can be used to quantify chemical diversity, in diversity picking, to represent large sections of chemical space, and in new clustering algorithms, which can in turn be applied to the study of protein folding landscapes.
Monday
3754410 - Reverse fingerprints, virtual pockets, and high-throughput virtual screening. Oh my!
11:15am - 11:40am USA / Canada - Central - August 22, 2022 | Location: Henry Clarke (Marriott Marquis Chicago)
Division: [CINF] Division of Chemical Information
Session Type: Oral - Hybrid
A mutual information-based activity labeling and scoring approach to reverse fingerprint (RF) analysis is introduced [1]. Starting with a series of ligands, only described in 2D, and each with known biological activity, the RF approach allows for the elucidation of important molecular structural motifs (as related to activity) along with the determination of a consensus pharmacophore; the consensus pharmacophore includes both atom-centered and projected features.

It is possible to combine this information, along with ligand shape and excluded volume features (derived from conformational sampling), to derive a "virtual pocket". This virtual pocket represents a hypothesis for how the ligands may orient in a bound state, and for what the binding pocket might look like around the bound ligands. This virtual pocket can be used for both retrospective analysis (to rationalize known structure-activity relationships), or converted to a 3D pharmacophore query that can be used to search databases for new active molecules in a vHTS context.

Following a concise description of the virtual pockets methodology described above, a series of illustrative examples will be presented, describing the performance of the vHTS experiments against a variety of pharmaceutically relevant targets.

Monday
Closing Remarks
11:40am - 11:45am USA / Canada - Central - August 22, 2022 | Location: Henry Clarke (Marriott Marquis Chicago)
Division: [CINF] Division of Chemical Information
Session Type: Oral - Hybrid

Computational Chemogenomics: Bridging Chemical and Biological Spaces:
02:00pm - 05:20pm USA / Canada - Central - August 22, 2022 | Location: Henry Clarke (Marriott Marquis Chicago)
Dr. Jose L Medina-Franco, Organizer, Presider
Division: [CINF] Division of Chemical Information
Session Type: Oral - Hybrid
Division/Committee: [CINF] Division of Chemical Information

The symposium will present the progress on the development and application of computational methods to characterize, explore and expand the chemical space and biological spaces and their intersection. The methods and practical applications in computational chemogenomics include but are not limited to artificial intelligence, data mining and visualization, and library design. Applications of virtual screening and inverse virtual screening (target fishing) for drug repurposing and multi-target drug discovery are welcome.

Monday
Introductory Remarks
02:00pm - 02:05pm USA / Canada - Central - August 22, 2022 | Location: Henry Clarke (Marriott Marquis Chicago)
Division: [CINF] Division of Chemical Information
Session Type: Oral - Hybrid

Monday
3736143 - A systematic chemogenomic analysis of the known and predicted biological activities of food compounds
02:05pm - 02:30pm USA / Canada - Central - August 22, 2022 | Location: Henry Clarke (Marriott Marquis Chicago)
Division: [CINF] Division of Chemical Information
Session Type: Oral - Hybrid
A large experimental effort is currently performed to identify the biological mechanisms of action of food compounds from a molecular point of view, in order to find novel nutraceuticals and scaffolds for drug design, as well as to rationalize the beneficial or harmful effects of foods on human health. Caffeine is paradigmatic in this respect, as it has been largely studied and applied as template to develop adenosine receptor antagonists, cyclic nucleotide phosphodiesterase inhibitors, and other biological targets. The identification of molecular mechanisms of action of food compounds is typically performed through biochemical and/or biological (cellular) assays directed towards different biological targets (typically proteins). However, the targets to test are frequently selected in a more or less ad hoc basis, instead of being based on a systematic, evidence-based ranking of the whole set of targets in the human genome. In this respect, the cheminformatic approach here described uses chemogenomic models trained on chemobiological databases to predict food compound-target interactions based solely on the compound structure. This would provide cheap and fast approach to prioritize the interactions to test. Our group is interested in the structure-activity analysis and modeling of food compounds as sources for the design of novel chemotypes for drugs and nutraceutics, using cheminformatic methods from the drug discovery field. In this regard, here we conduct a systematic analysis of food compound vs human target interactions. In a first stage, we identify and analyze all the available experimental data for food compounds present in chemobiological databases. Then, we use cheminformatic statistical models to predict interactions with human targets, and analyze the results as well. From here, a series of patterns emerge that are specific for these group of compounds, both from the point of view of the structure of the compound, as from the point of view of the target classes. These will be described in this presentation.
Monday
3750458 - De novo molecular design with genetic algorithms using alvaBuilder
02:30pm - 02:55pm USA / Canada - Central - August 22, 2022 | Location: Henry Clarke (Marriott Marquis Chicago)
Division: [CINF] Division of Chemical Information
Session Type: Oral - Hybrid
Designing molecules with certain desirable properties is a topic, also referred to as de novo molecular generation, of interest to many research fields. It can be seen as a combinatorial optimization problem in which the goal is to explore the chemical space in search of molecules for further study or synthesis. The estimated size of the chemical space of even drug-like molecules alone is so huge that is not possible to explore it all. We present an approach, implemented in a software tool called alvaBuilder, which constraints the search to molecules that can be composed of a set of fragments. Such fragments are identified by splitting the molecules contained in a given training data set. The exploration is then performed by assembling the aforementioned fragments to generate candidate molecules. Each candidate is associated to a score based on the desired properties defined by the researcher. By using Genetic Algorithms, alvaBuilder generates a final pool with the molecules that have the best score. The Genetic Algorithms take inspiration by Darwinian natural selection assuming that the best fitted members of a population of molecules survive and improvements can be achieved by mutating and recombining their genes, i.e. their molecular fragments. The purpose of this talk is to present how to perform de novo molecular generation using alvaBuilder with some practical sample cases to design molecules with certain phisicochemical properties.
Monday
3741397 - Common Scaffold Visualizer (CSViz): A computational framework to enable interactive analysis and design of multi-targeted kinase inhibitors
02:55pm - 03:20pm USA / Canada - Central - August 22, 2022 | Location: Henry Clarke (Marriott Marquis Chicago)
Division: [CINF] Division of Chemical Information
Session Type: Oral - Hybrid
Agents targeting multiple proteins are known to circumvent efficacy and drug resistance-related limitations in comparison to highly selective drugs. On the other hand, these therapeutics often have unintended off-target interactions giving rise to adverse drug effects. Rational design of agents with targeted polypharmacology requires a detailed understanding of both the chemical space of the targets under investigation and the landscape of the anti-targets. Though there exists a plethora of computer-aided drug design (CADD) tools geared towards creating selective inhibitors, the rational design of multi-targeting agents calls for a novel suite of tools that best accommodates the polypharmacological rationale.

To address this unmet need, we present CSViz, a computational tool that interactively analyzes the chemogenomic landscape of targets and anti-targets to 1) identify scaffolds shared across targets and 2) exclude scaffolds implicated in anti-target binding. CSViz thereby provides a knowledge-driven compound moiety for designing multi-target inhibitors. Given a set of targets and their inhibitors, CSViz estimates pairwise compound similarity and finds the maximum common substructure (MCS). These MCSs are first filtered based on connectivity and ring-based criteria, and later subjected to topological-based clustering. Representative MCSs are identified for each cluster and their enrichment scores are estimated. The resulting MCSs, along with their enrichment scores, enable the identification of initial scaffolds for the design of ligands with multi-targeting activity.

The present case study focuses on identifying polypharmacological agents for kinase targets p38α and EGFR. Representative MCSs shared between inhibitors of p38α and EGFR were identified (Fig. 1). These MCSs were used as seed molecules to virtually screen various vendor libraries. The compounds were found to show favorable docking poses amongst the selected targets. We anticipate that this tool will accelerate the discovery of kinase inhibitors with targeted polypharmacology.
Fig. 1. MCS shared between p38α and EGFR kinases: The MCS-1 shared between EGFR and MAPK11 (p38α) identified through CSViz is highlighted in both compounds.

Fig. 1. MCS shared between p38α and EGFR kinases: The MCS-1 shared between EGFR and MAPK11 (p38α) identified through CSViz is highlighted in both compounds.


Monday
3743743 - Computational pipeline to identify potential drug targets and interacting chemotypes in SARS-CoV-2
03:20pm - 03:45pm USA / Canada - Central - August 22, 2022 | Location: Henry Clarke (Marriott Marquis Chicago)
Division: [CINF] Division of Chemical Information
Session Type: Oral - Hybrid
Minimizing the human and economic costs of the COVID-19 pandemic and of future pandemics requires the ability to develop and deploy effective treatments for novel pathogens as soon as possible after they emerge. Yet drug discovery is a notoriously slow process. An accurate and streamlined computational method for predicting potential drug target sites could reduce the time and cost of experimental screenings and serve a critical purpose in accelerating drug development.
To this end, we introduce a unique, computational pipeline for the rapid identification and characterization of binding sites in the proteins of novel viruses as well as the core chemical components with which these sites interact. We combine molecular-level structural modeling of proteins with data science clustering and cheminformatic techniques in a computationally efficient manner. We leverage the PDBspheres method, based on experimental crystal structures from PDB, to implement structural modeling of SARS-CoV-2 proteins and produce protein-ligand binding data. Using this data, we first group protein residues based on their ligand contacts and select groups that form relevant binding sites. We then group the ligands that bind to each site based on their chemical structure and function, revealing the site-specific, key chemical features involved in binding.
Similarities between our results, experimental data, and other computational studies provide support for the effectiveness of our predictive framework. We find that the composition of structural models for an individual binding site may be used as a simple heuristic for determining which binding sites will make more effective drug targets. While we present here a demonstration of our tool on SARS-CoV-2, our process is generalizable and can be applied to any new virus, as long as either experimentally solved structures for its proteins are available or sufficiently accurate homology models can be constructed.

Monday
Intermission
03:45pm - 04:00pm USA / Canada - Central - August 22, 2022 | Location: Henry Clarke (Marriott Marquis Chicago)
Division: [CINF] Division of Chemical Information
Session Type: Oral - Hybrid

Monday
3741929 - Autonomous molecular design and targeted transformations of small molecule space into specific protein kinase inhibitors using hierarchical perturbation-based machine learning approaches
04:00pm - 04:25pm USA / Canada - Central - August 22, 2022 | Location: Henry Clarke (Marriott Marquis Chicago)
Division: [CINF] Division of Chemical Information
Session Type: Oral - Hybrid
Development of machine learning solutions for targeted exploitation of enormous chemical space and objective-guided molecular design are paramount in modern biomedical research and have gained a significant momentum. We present mathematical formalisms and implementation of generative machine learning strategies for objective-guided molecular design and cross-domain transformation of small molecules to the domain of selective probes of protein kinases. We developed a portfolio of machine learning approaches that include Generative Adversarial Networks (GAN) for objective-guided molecular design of specific chemical probes targeting the ABL and SRC kinase; (b) de novo generative machine learning approach with Bayesian optimization and chemical perturbation; (c) generative perturbation learning for drug repurposing into targeted specific kinase probes. For generation of a comprehensive dataset of kinase chemical probes used in machine learning training, we consolidate prioritize and integrate known kinase inhibitors from 10 major kinase families along with small molecule databases including ZINC and the GDB-17. After encoding and mapping molecules in a continuous latent space representation, we develop a robust chemical feature-based machine learning predictor of kinase inhibition likelihood that is used in perturbation-based transformation of small molecules. The developed perturbation-based generative design approaches allow for an efficient manipulation and transformation of chemical space into predefined classes of specific kinase probes. By combining molecular perturbation design with the kinase inhibition likelihood analysis and similarity assessments, we demonstrate how autonomous molecular design strategy can morph inhibitors of LCK, ABL1, and EGFR kinases into selective chemical probes of SRC kinase. We will discuss how integration of experiment-informed protein modeling, network science and generative machine learning approaches could lead to a new general strategy for autonomous molecular design and validation of targeted and allosteric chemical probes with applications to families of protein kinases and molecular chaperones.
Monday
3750710 - Crafting opiates: Learning from the opioid chemspace and incorporating a molecular spark
04:25pm - 04:50pm USA / Canada - Central - August 22, 2022 | Location: Henry Clarke (Marriott Marquis Chicago)
Division: [CINF] Division of Chemical Information
Session Type: Oral - Hybrid
Biased activation of G-protein-coupled receptors (GPCRs) is shifting drug discovery efforts and appears promising for the development of safer drugs. The most effective analgesics to treat acute pain are agonists of the mu-opioid receptor (m-OR), a member of the GPCR superfamily. However, their therapeutic use requires close medical monitoring to diminish the risk of severe adverse effects. The G-protein-biased agonists of m-ORs have shown safer therapeutic profiles than non-biased ligands. In this talk, we will present the analysis of the chemical space covered by opioid ligands, a fingerprint codification of biased agonisim, and strategies for the modification of the pharmacological properties.
Monday
3736504 - TidyMass: An object-oriented reproducible analysis framework for LC-MS data
04:50pm - 05:15pm USA / Canada - Central - August 22, 2022 | Location: Henry Clarke (Marriott Marquis Chicago)
Dr. Xiaotao Shen, Presenter
Division: [CINF] Division of Chemical Information
Session Type: Oral - Hybrid
LC-MS-based metabolomics has been an important tool in research. However, analyzing metabolomics data is still challenging. To overcome the challenges, the community has developed numerous tools. However, some limitations exist. Commercial tools only work on the associated instrument platform, online/GUI tools cannot take advantage of the cluster computational power, open-source tools typically follow limited parts of the whole workflow and have no uniform format for data input.
Therefore, we developed tidymass, which is a comprehensive computational framework (Figure 1). Tidymass was developed based on the following strategies.
1) Cross-platform utility. It can be installed on all platforms, so can utilize the computational power of the cluster.
2) Uniform, shareable, traceable, and reproducible. A uniform data format “mass_dataset” class is designed to store metabolomics data and processing information; all the packages in tidymass use the “mass_dataset” as the input data format (Figure 2). This uniform data structure makes it easier to share data. In addition, all the parameters are stored in this class, which makes it is possible to trace the processing parameters used. This provides a highly reproducible and robust analytical framework.
3) Flexible and extensible. Tidymass is a collection of multiple R packages, in which the different packages correspond to different steps of the workflow (Figure 1), which makes it easy for the user to find appropriate functions, and for developers to debug and extend. This modular design makes tidymass a highly flexible tool.

Monday
Concluding Remarks
05:15pm - 05:20pm USA / Canada - Central - August 22, 2022 | Location: Henry Clarke (Marriott Marquis Chicago)
Division: [CINF] Division of Chemical Information
Session Type: Oral - Hybrid