Aaron Chuah

Flow cytometry is a powerful analytical technique that generates high-dimensional single-cell data widely used in biological research and clinical diagnostics. This survey traces the evolution of flow cytometry analysis from traditional manual gating to modern algorithmic approaches and explores emerging deep-learning (DL) paradigms. We delineate three distinct analysis paradigms: traditional (manual gating), modern (algorithm-assisted population identification), and future (direct cell-to-sample DL models). While the modern paradigm employs sophisticated computational methods to identify cell populations and derive features, the emerging future paradigm bypasses discrete population identification entirely, utilizing DL to directly connect cellular information to sample-level insights. We discuss key technical challenges including batch effects, panel variations, and data availability that currently limit widespread adoption of more advanced analytical approaches. The survey also examines promising solutions such as transfer learning, data augmentation, and self-supervised learning techniques. As flow cytometry technology continues to advance, particularly with spectral cytometry enabling higher-dimensional analysis, these computational methods will become increasingly essential for extracting maximum biological insight from complex datasets. The development of standardized data repositories, robust batch correction tools, and comprehensive benchmarking frameworks will be crucial for realizing the full potential of these advanced analytical approaches.

Summary: Complex tabular datasets comprising many diverse features can require specific expertise to interpret, posing a barrier to researchers with minimal data science experience. EDAmame is an interactive tool that simplifies initial analysis and visualization of these datasets, providing insights into data quality and feature relationships. By leveraging open-source machine learning frameworks in R, EDAmame allows researchers to perform effective exploratory data analysis without command-line or coding requirements. Availability and implementation: A limited online version can be accessed at https://edamame.org.au/ or can be downloaded from https://doi.org/10.5281/zenodo.15356492. The app is developed in R Shiny and implements tidyverse and tidymodels packages.

This study compares the relative utility of deep learning models as automated phenotypic classifiers, built with features of peripheral blood cell populations assayed with flow cytometry. We report surprisingly good sensitively for the task of distinguishing male from female mice solely from attributes of cells present in peripheral blood samples. This novel application of cytometry technology is easily validated and represents an example use that is not reliably possible with current visual analysis techniques. It serves as a useful test case with known ground truth labels of a subtle cellular phenotype and indicates the potential of this approach to more biologically meaningful phenotypes. We evaluate three different deep neural network architectures that vary in how the high-dimensional flow cytometry data is processed as input: 1) a convolutional neural network applied to the raw, per-cell marker intensity values, 2) a multi-layer perceptron taking per-channel histograms representing the range of intensities for a given marker, and 3) a multi-layer perceptron using a hypervoxel representation of the set of marker intensities for each cell in a given input sample. Our results show the histogram-based multi-layer perceptron achieves the highest classification accuracy of 91% on 2300 mouse blood samples for a mutant phenotype classification task. This indicates the value of deep learning models for detection of subtle signals that are not readily observed in high dimensions by visual analysis. These results demonstrate the effectiveness of deep learning for automated analysis of flow cytometry data on a challenging and biologically complex classification task.

Missense mutations that disrupt protein structural stability are a common pathogenic mechanism in human genetic disease. Here, we quantify potential disruption of protein stability due to amino acid substitution and show that functionally constrained proteins are less susceptible to large mutational changes in stability. Mechanistically, this relates to greater intrinsic disorder among constrained proteins and to increased B-factors in the ordered regions of constrained proteins. This phenomenon means that constrained proteins exhibit smaller stability effects due to missense mutations, and partly explains why overtransmission of pathogenic missense variation is less prevalent in genetic disorders characterised by protein truncations. We show that the most functionally constrained proteins are depleted of both destabilising and overly-stabilising amino acid variation in disease-free populations. Despite this, amino acid substitutions with large stability effects in functionally constrained proteins are still highly prevalent among pathogenic human genetic variation. Importantly, we observe that there are approximately five times more missense variants with large stability effects than there are unambiguous loss-of-function mutations. Missense variants with disruption of stability effects recapitulate the per-gene patterns of functional constraint observed with protein truncating loss-of-function variation, yet their relative abundance abrogates difficulties encountered when estimating functional constraint for the shortest human genes.

Systemic lupus erythematosus (SLE) is a heterogeneous autoimmune disease with a clear genetic component. While most SLE patients carry rare gene variants in lupus risk genes, little is known about their contribution to disease pathogenesis. Amongst them, SH2B3—a negative regulator of cytokine and growth factor receptor signaling—harbors rare coding variants in over 5% of SLE patients. Here, we show that unlike the variant found exclusively in healthy controls, SH2B3 rare variants found in lupus patients are predominantly hypomorphic alleles, failing to suppress IFNGR signaling via JAK2-STAT1. The generation of two mouse lines carrying patients’ variants revealed that SH2B3 is important in limiting the number of immature and transitional B cells. Furthermore, hypomorphic SH2B3 was shown to impair the negative selection of immature/transitional self-reactive B cells and accelerate autoimmunity in sensitized mice, at least in part due to increased IL-4R signaling and BAFF-R expression. This work identifies a previously unappreciated role for SH2B3 in human B cell tolerance and lupus risk.

Systemic lupus erythematosus (SLE) is a heterogeneous autoimmune disease, with a clear genetic component. While most SLE patients carry rare gene variants in lupus risk genes, little is known about their contribution to disease pathogenesis. Amongst them, SH2B3 - a negative regulator of cytokine and growth factor receptor signaling – harbors rare coding variants in over 5% of SLE patients. Here we show that unlike the variant found exclusively in healthy controls, most SH2B3 rare variants found in lupus patients are predominantly hypomorphic alleles. Generation of two mouse lines carrying variants orthologous to those found in patients revealed SH2B3 is important to limit the numbers of immature and transitional B cells. Furthermore, hypomorphic SH2B3 was shown to impair negative selection of immature/transitional self-reactive B cells and accelerate autoimmunity in sensitized mice, at least in part due to increased IL-4R signaling and BAFF-R expression. This work identifies a previously unappreciated role for SH2B3 in human B cell tolerance and lupus risk. Summary: Zhang et al . reveal a role for hypomorphic SH2B3 in lupus risk. The study shows rare and damaging variants identified in lupus patients enable breach of B cell immune tolerance checkpoints and suggests involvement for dysregulated IL-4R signaling and BAFF-R expression.

Missense mutations that change protein stability are strongly associated with human genetic disease. With the recent availability of predicted structures for all human proteins generated using the AlphaFold2 prediction model, genome-wide assessment of the stability effects of genetic variation can, for the first time, be easily performed. This facilitates the interrogation of personal genetic variation for potentially pathogenic effects through the application of stability metrics. Here, we present a novel tool to prioritize variants predicted to cause strong instability in essential proteins. We show that by filtering by ΔΔG values and then prioritizing by StabilitySort Z-scores, we are able to more accurately discriminate pathogenic, protein-destabilizing mutations from population variation, compared with other mutation effect predictors. StabilitySort is available as a web service (https://www.stabilitysort.org), as a data download for integration with other tools (https://www.stabilitysort.org/download) or can be deployed as a standalone system from source code (https://gitlab.com/baaron/StabilitySort). Supplementary data are available at Bioinformatics online.

Background and Objectives: Chronic kidney disease progression to ESKD is associated with a marked increase in mortality and morbidity. Its progression is highly variable and difficult to predict. Methods: This is an observational, retrospective, single-centre study. The cohort was patients attending hospital and nephrology clinic at The Canberra Hospital from September 1996 to March 2018. Demographic data, vital signs, kidney function test, proteinuria, and serum glucose were extracted. The model was trained on the featurised time series data with XGBoost. Its performance was compared against six nephrologists and the Kidney Failure Risk Equation (KFRE). Results: A total of 12,371 patients were included, with 2,388 were found to have an adequate density (three eGFR data points in the first 2 years) for subsequent analysis. Patients were divided into 80%/20% ratio for training and testing datasets. ML model had superior performance than nephrologist in predicting ESKD within 2 years with 93.9% accuracy, 60% sensitivity, 97.7% specificity, 75% positive predictive value. The ML model was superior in all performance metrics to the KFRE 4- and 8-variable models. eGFR and glucose were found to be highly contributing to the ESKD prediction performance. Conclusions: The computational predictions had higher accuracy, specificity and positive predictive value, which indicates the potential integration into clinical workflows for decision support.

Summary: Missense mutations that change protein stability are strongly associated with human inherited genetic disease. With the recent availability of predicted structures for all human proteins generated using the AlphaFold2 prediction model, genome-wide assessment of the stability effects of genetic variation can, for the first time, be easily performed. This facilitates the interrogation of personal genetic variation for potentially pathogenic effects through the application of stability metrics. Here, we present a novel algorithm to prioritise variants predicted to strongly destabilise essential proteins, available as both a standalone software package and a web-based tool. We demonstrate the utility of this tool by showing that at values of the Stability Sort Z-score above 1.6, pathogenic, protein-destabilising variants from ClinVar are detected at a 58% enrichment, over and above the destabilising (but presumably non-pathogenic) variation already present in the HapMap NA12878 genome. Availability and Implementation: StabilitySort is available as both a web service ( http://130.56.244.113/StabilitySort/ ) and can be deployed as a standalone system ( https://gitlab.com/baaron/StabilitySort ). Contact: Dan.Andrews@anu.edu.au

Objective: To compare the frequency of rare variants in genes of the pathophysiologically relevant endosomal Toll-like receptor (eTLR) pathway and any quantifiable differences in variant rarity, predicted deleteriousness, or molecular proximity in patients with systemic lupus erythematosus (SLE) and healthy controls. Patients and methods: 65 genes associated with the eTLR pathway were identified by literature search and pathway analysis. Using next generation sequencing techniques, these were compared in two randomised cohorts of patients with SLE (n = 114 and n = 113) with 197 healthy controls. Genetically determined ethnicity was used to normalise minor allele frequencies (MAF) for the identified genetic variants and these were then compared by their frequency: rare (MAF < 0.005), uncommon (MAF 0.005–0.02), and common (MAF >0.02). This was compared to the results for 65 randomly selected genes. Results: Patients with SLE are more likely to carry a rare nonsynonymous variant affecting proteins within the eTLR pathway than healthy controls. Furthermore, individuals with SLE are more likely to have multiple rare variants in this pathway. There were no differences in rarity, Combined Annotation Dependent Depletion (CADD) score, or molecular proximity for rare eTLR pathway variants. Conclusions: Rare non-synonymous variants are enriched in patients with SLE in the eTLR pathway. This supports the hypothesis that SLE arises from several rare variants of relatively large effect rather than many common variants of small effect.