Development of a Ranking System for Bayesian Kinase Network Models
Alex Joyce, The University of Toledo, United States
GLBIO 2021
Accurate modeling of kinase networks is an increasingly relevant topic in current research due to increased interest in kinase-kinase interactions. An example of a modeling system is KINNET, an R package developed by Ali Imami which uses Bayesian Networks to model these interactions using kinomic data. However, the number of potential networks increases exponentially with more kinases. As such, it is important to develop a method to select the optimal kinase when presented with the choice between multiple kinases with similar scores. Here, a ranking system is presented to add weight to kinases based on two parameters: the number of interactions that a kinase participates in and the knowledge base available for that particular kinase.
This two-parameter approach aims to balance the complexity that a kinase adds to the network with how much information is available, allowing for the selection or rejection of potential nodes depending on the needs of the user. In addition, the ability to switch between “dark” and “light” kinases, which are those with a lack or abundance of information respectively, has also been developed, which may enable the discovery of novel kinase interactions.
Analysis of CNS disorders using active kinome profiles
*A. JOYCE, K. ALGANEM, A. S. IMAMI, J. CREEDEN, A. HAMOUD, W. RYAN, R. SHUKLA, R. MCCULLUMSMITH;
Neurosci. & Neurolog. Disorders, Univ. of Toledo Col. of Med., Toledo, OH
SFN 2021
Central nervous system (CNS) disorders such as AD (Alzheimer’s dementia) and schizophrenia are often caused by a variety of pathophysiological processes, many of which involve abnormal protein kinase signaling. The active kinome is the set of kinases actively regulating cellular functions through phosphorylation. The active kinome can be profiled using protein kinase activity arrays. Using these profiles, we can empirically determine the phosphorylation state of peptide targets and find commonalities between the mechanisms of CNS disorders. To determine how the active kinome regulates the mechanisms underlying CNS disorders, we will first identify peptides that are differentially phosphorylated in AD and schizophrenia postmortem brain samples versus control subjects. We will then use deconvolution methods to identify the top dysregulated kinasesTo determine how the active kinome regulates the mechanisms underlying CNS disorders, we will first validate kinases that have been implicated in CNS disorders by our kinome array profile. We will identify peptides on the array that are differentially phosphorylated in Alzhemier’s dementia and schizophrenia. Next, techniques such as BLAST will be used to assign these peptides to their respective proteins and identify homologous sequences mapped to other proteins. We will use perform an in-silico data exploration of databases such as BrainAtlas to determine the regional and cell level protein expression patterns of both the kinases and targets in the brain. We will use Enrichr to perform a pathway analysis on the final list of proteins to determine their functional implications. To have a more thorough understanding of the effects of the active kinome and the downstream effects of its protein targets, we will compare the two sets of pathways generated from AD (Alzheimers) and schizophrenia. Finally, we will curate publicly available CNS datasets from phosphoproteomic databases of CNS datasets disorders and explore the overlapping pathways using these data sets to get a more complete picture of the commonalities between these disorders regarding dysregulated kinases. We believe that this dysregulation plays an important role in CNS disorders and that common pathways regarding these dysregulated kinases could provide a unique insight into the mechanisms of CNS disorders and their methods of treatment.
Probabilistic validation for targeted proteomics using parallel reaction monitoring
Alex W. Joyce,1,2 Yameng Wu,1 and Brian C. Searle1,2,3
1 Pelotonia Institute for Immuno-Oncology, Comprehensive Cancer Center The Ohio State University
2 Department of Biomedical Informatics, The Ohio State University Medical Center
3 Department of Chemistry and Biochemistry, The Ohio State University
OMSS 2023
Mass spectrometry is a powerful method for conducting high-throughput proteomic experiments. While global methods such as data-dependent acquisition (DDA) are effective for large-scale peptide identification, they are often unreliable for analyzing specific proteins of interest due to reproducibility issues caused by stochastic sampling. Targeted methods such as Selected Reaction Monitoring (SRM) and Parallel Reaction Monitoring (PRM) serve as alternatives that are more suitable for such tasks because they regularly sample targeted peptides regardless of the measured signal. A commonly used metric for validating proteomics experiments is the false discovery rate (FDR), which estimates the number of false positive matches in a dataset. While FDR thresholding is effective for evaluating results globally, targeted proteomics experiments could benefit from estimating a posterior error probability for each targeted peptide. SRM measurements are performed only on selected transitions that are tailored to targeted peptides, which limits the measurement size. On the other hand, PRM experiments computationally extract transition ions from full scan MS/MS spectra, which allows for other peptides that fall in the same retention time and precursor isolation window to also be measured. These additional “bonus” peptides can be used to derive distributions and statistics that help determine posterior error probabilities for individual peptides. Here, we introduce a method for determining these probabilities from the scoring distributions of “bonus” peptides identified in PRM. We expect that this will be a valuable method for evaluating and validating the accuracy of peptides measured by PRM.
Exploring biofluid protein expression with CATalog, an interactive proteomics dashboard
Alex W. Joyce,1 Katelyn B. Brusach,2 Jessica M. Quimby,2 Brian C. Searle1,3
1Department of Biomedical Informatics, The Ohio State University Medical Center, Columbus, OH, United States
2Veterinary Clinical Sciences, The Ohio State University College of Veterinary Medicine, Columbus, OH, United States
3Department of Chemistry and Biochemistry, The Ohio State University, Columbus, OH, United States
OMSS 2024
Biofluids contain many different potential biomarkers, providing valuable insight into both normal and aberrant biological processes. However, it is not always apparent which biofluid to choose when evaluating biomarker candidates as specific proteins are detected at different concentrations among various biofluids. Pursuing the wrong biofluid can result in limited detection or results with lower significance. To aid in the process of selecting the correct biofluid for investigation, we have developed CATalog, an interactive application written in R using the Shiny framework. CATalog displays baseline relative intensities of identified proteins from paired urine, serum, and plasma samples measured using mass spectrometry-based proteomics. These baseline intensities of healthy individuals aids biomarker discovery by providing a reference point to determine abnormal protein levels in disease. Additionally, users of this application have the option to filter out proteins with highest relative intensities in any given biofluid, view boxplots to visualize data, identify outliers that correspond to demographic factors, and obtain gene ontology (GO) annotations to be viewed for every protein. The integration of GO annotations into the database allows the end user to quickly assess both function and subcellular localization of a given biomarker, allowing the application to assist in both biofluid and biomarker selection. We demonstrate this software tool using a core database of feline proteomic data collected from nine healthy cats. We believe that this application is a useful tool for both biofluid selection and biomarker exploration in the field of proteomics.
Alex W. Joyce 