The role of saliency maps in enhancing ophthalmologists' trust in artificial intelligence models.

PURPOSE: Saliency maps (SM) allow clinicians to better understand the opaque decision-making process in artificial intelligence (AI) models by visualising the important features responsible for predictions. This ultimately improves interpretability and confidence. In this work, we review the use case for SMs, exploring their impact on clinicians' understanding and trust in AI models. We use the following ophthalmic conditions as examples: (1) glaucoma, (2) myopia, (3) age-related macular degeneration, and (4) diabetic retinopathy. METHOD: A multi-field search on MEDLINE, Embase, and Web of Science was conducted using specific keywords. Only studies on the use of SMs in glaucoma, myopia, AMD, or DR were considered for inclusion. RESULTS: Findings reveal that SMs are often used to validate AI models and advocate for their adoption, potentially leading to biased claims. Overlooking the technical limitations of SMs, and the conductance of superficial assessments of their quality and relevance, was discerned. Uncertainties persist regarding the role of saliency maps in building trust in AI. It is crucial to enhance understanding of SMs' technical constraints and improve evaluation of their quality, impact, and suitability for specific tasks. Establishing a standardised framework for selecting and assessing SMs, as well as exploring their relationship with other reliability sources (e.g. safety and generalisability), is essential for enhancing clinicians' trust in AI. CONCLUSION: We conclude that SMs are not beneficial for interpretability and trust-building purposes in their current forms. Instead, SMs may confer benefits to model debugging, model performance enhancement, and hypothesis testing (e.g. novel biomarkers).

Assessment of artificial intelligence to detect gasoline in fire debris using HS-SPME-GC/MS and transfer learning.

Due to the complex nature of the chemical compositions of ignitable liquids (IL) and the interferences from fire debris matrices, interpreting chromatographic data poses challenges to analysts. In this work, artificial intelligence (AI) was developed by transfer learning in a convolutional neural network (CNN), GoogLeNet. The image classification AI was fine-tuned to create intelligent classification systems to discriminate samples containing gasoline residues from burned substrates. All ground truth samples were analyzed by headspace solid-phase microextraction (HS-SPME) coupled with a gas chromatograph and mass spectrometer (GC/MS). The HS-SPME-GC/MS data were transformed into three types of image presentations, that is, heatmaps, extracted ion heatmaps, and total ion chromatograms. The abundance and mass-to-charge ratios of each scan were converted into image patterns that are characteristic of the chemical profiles of gasoline. The transfer learning data were labeled as "gasoline present" and "gasoline absent" classes. The assessment results demonstrated that all AI models achieved 100 ± 0% accuracy in identifying neat gasoline. When the models were assessed using the spiked samples, the AI model developed using the extracted ion heatmap obtained the highest accuracy rate (95.9 ± 0.4%), which was greater than those obtained by other machine learning models, ranging from 17.3 ± 0.7% to 78.7 ± 0.7%. The proposed work demonstrated that the heatmaps created from GC/MS data can represent chemical features from the samples. Additionally, the pretrained CNN models are readily available in the transfer learning workflow to develop AI for GC/MS data interpretation in fire debris analysis.

Use of eye-gaze technology feedback by assistive technology professionals: findings from a thematic analysis.

Purpose: Eye-gaze technology offers professionals a range of feedback tools, but it is not well understood how these are used to support decision-making or how professionals understand their purpose and function. This paper explores how professionals use a variety of feedback tools and provides commentary on their current use and ideas for future tool development.Methods and Materials: The study adopted a focus group methodology with two groups of professional participants: those involved in the assessment and provision of eye-gaze technology (n = 6) and those who interact with individuals using eye-gaze technology on an ongoing basis (n = 5). Template analysis was used to provide qualitative insight into the research questions.Results: Professionals highlighted several issues with existing tools and gave suggestions on how these could be made better. It is generally felt that existing tools highlight the existence of problems but offer little in the way of solutions or suggestions. Some differences of opinion related to professional perspective were highlighted. Questions about automating certain processes were raised by both groups.Conclusions: Discussion highlighted the need for different levels of feedback for users and professionals. Professionals agreed that current tools are useful to identify problems but do not offer insight into potential solutions. Some tools are being used to draw inferences about vision and cognition which are not supported by existing literature. New tools may be needed to better meet the needs of professionals and an increased understanding of how existing tools function may support such development.

Professionals sometimes make use of feedback tools to infer the cognitive and/or visual abilities of users, although the tools are not designed or validated for these purposes, and the existing literature does not support this.Some eye-gaze feedback tools are perceived as a "black box", leaving professionals uncertain as to how to usefully interpret and apply the outputs.There is an opportunity to improve tools that provide feedback on how well an eye-gaze system is working or how effectively a user can interact with this technology.Professionals identified that tools could be better at offering potential solutions, rather than simply identifying the existence of problems.

Effects of various cross-linked collagen scaffolds on wound healing in rats model by deep-learning CNN.

Scar tissue is connective tissue formed on the wound during the wound-healing process. The most significant distinction between scar tissue and normal tissue is the appearance of covalent cross-linking and the amount of collagen fibers in the tissue. This study investigates the efficacy of four types of collagen scaffolds in promoting wound healing and regeneration in a Sprague-Dawley murine model-the histomorphology analysis of collagen scaffolds and developing a deep learning model for accurate tissue classification. Four female rats (n = 24) groups received collagen scaffolds prepared through physical and chemical crosslinking. Wound healing progress was evaluated by monitoring granulation tissue formation, collagen matrix organization, and collagen fiber deposition, with histological scoring for quantification-the EDC and HA groups demonstrated enhanced tissue regeneration. The EDC and HA groups observed significant differences in wound regeneration outcomes. Deep-learning CNN models with data augmentation techniques were used for image analysis to enhance objectivity. The CNN architecture featured pre-trained VGG16 layers and global average pooling (GAP) layers. Feature visualization using Grad-CAM heatmaps provided insights into the neural network's focus on specific wound features. The model's AUC score of 0.982 attests to its precision. In summary, collagen scaffolds can promote wound healing in mice, and the deep learning image analysis method we proposed may be a new method for wound healing assessment.

MRI features of neuronal intranuclear inclusion disease, combining visual and quantitative imaging investigations.

OBJECTIVE: To observe the radiological characteristics of Neuronal Intranuclear Inclusion Disease (NIID) on lesion locations and diffusion property using quantitative imaging analysis. METHODS: Visual inspection and quantitative analyses were performed on MRI data from 31 retrospectively included patients with NIID. Frequency heatmaps of lesion locations on T2WI and DWI were generated using voxel-wise analysis. Gray matter volume (GMV), white matter volume (WMV) and diffusion property of apparent diffusion coefficient (ADC) values of patients were voxel-wisely compared with healthy controls. Moreover, the ADC values within the DWI-detected lesion were compared with those within the adjacent cortical gray matter and white matter. Voxel-based lesion symptom mapping (VLSM) techniques, were used to determine the relationship between DWI lesion location and disease durations. RESULTS: By visual inspection on the imaging findings, we proposed an "cockscomb flower sign" for describing the radiological feature of DWI hyperintensity within the corticomedullary junction. A "T2WI-DWI mismatch of spatial distribution" pattern was also revealed with visual inspection and frequency heatmaps, for describing the feature of a wider lesion distribution covering white matter shown on T2WI than that on DWI. Voxel-based morphometry comparison revealed that wildly reduced GMV and WMV, both the lesion areas detected by DWI and T2WI demonstrated ADC increase in patients. Furthermore, the ADC values within the DWI-detected lesion were intermediate between the adjacent cortex and the deep white matter with highest ADC. VLSM analysis revealed that frontal lobe, parietal lobe and internal capsule damage were associated with higher NIID durations. CONCLUSION: NIID features with "cockscomb flower-like" DWI hyperintensity in area of corticomedullary junction, based on a "T2WI-DWI mismatch of spatial distribution" of lesion locations. The pathological substrate of corticomedullary junction hyperintensity on DWI, can not be explained as diffusion restriction. These typical radiological features of brain MRI would be helpful for diagnosis of NIID.

Secondary metabolites produced by four Colletotrichum species in vitro and on fruits of diverse olive cultivars.

This study was aimed to characterize the secondary metabolites produced by four Colletotrichum species, C. acutatum, C. gloeosporioides, C. godetiae and C. karsti, both in vitro, on potato dextrose agar (PDA) and oatmeal agar (OA), and during the infection process of fruits of four olive cultivars differing in susceptibility to anthracnose, 'Coratina' and 'Ottobratica', both susceptible, 'Frantoio' and 'Leccino', both resistant. The metabolites were extracted from axenic cultures after seven days incubation and from olives inoculated singularly with each Colletotrichum species, at three different times, 1, 3 and 7 days post inoculation (dpi). They were identified using the UHPLC-QTOF-MS analysis method. In total, as many as 45 diverse metabolites were identified. Only 10 metabolites were present in both fruits and axenic cultures while 19 were found exclusively on olives and 16 exclusively in axenic cultures. The identified metabolites comprised fatty acid, phenolics, pyrones, sterols, terpenes and miscellaneous compounds. Each Colletotrichum species produced a different spectrum of metabolites depending on the type of matrices. On artificially inoculated olives the severity of symptoms, the amount of fungal secondary metabolites and their number peaked 7 dpi irrespective of the cultivar susceptibility and the virulence of the Colletotrichum species.

Highlighting Shooting Opportunities in Football.

The purpose of the present study was to create a two-dimensional model which illustrates a landscape of shooting opportunities at goal during a competitive football match. For that purpose, we analysed exemplar attacking subphases of each team when the ball was in the last 30 m of the field. The player's positional data (x and y coordinates) and the ball were captured at 25 fps and processed to create heatmaps that illustrated the shooting opportunities that were available in the first and second half in different field areas. Moreover, the time that the shooting opportunities were available was estimated. Results show that in the observed match, most of the shooting opportunities lasted between 1 and 2 s, with only a few opportunities lasting more than 2 s. The shooting opportunities did not display a homogenous distribution over the field. The obtained heatmaps provide valuable and specific information about each team's shooting opportunities, allowing the identification of the most vulnerable areas. Additionally, the amount, duration, and location of the shooting opportunities have shown significant differences between teams. This customizable model is sensitive to the features of shooting opportunities and can be used in real-time video analysis for individual and collective performance analysis.

Opioid-related mortality: Dynamic temporal and spatial trends by drug type and demographic subpopulations, Massachusetts, 2005-2021.

BACKGROUND: Fatal opioid-related overdoses (OOD) present significant public health challenges. Intuitive and replicable analytical approaches are needed to inform targeted public health responses. METHODS: We obtained fatal OOD data for 2005-2021 from the Massachusetts Registry of Vital Records and Statistics. We conducted heatmap analyses to assess trends in fatal OOD rates per 100,000 residents, visualizing rates by death year and decedent age at one-year intervals, stratifying by race/ethnicity, sex, rurality, and involved substances. We calculated Getis-Ord Gi* statistics to identify spatial clusters of OOD rates. RESULTS: Among 20,774 fatal OODs, rates were higher among males, and highly variable by race/ethnicity, age group, and rurality. While fatal OOD rates increased in urban before rural communities, rates were higher in rural communities by 2018-2019. Stimulant-related fatal OODs were elevated in 2020 and 2021. Fatal OOD rates involving fentanyl and stimulants increased precipitously and simultaneously in the non-Hispanic Black population in 2020 and 2021, with a bimodal age distribution peaking among those in their 40s and 60s. Elevated rates among 30-to-60 year old Hispanic residents were largely tied to synthetic opioids from 2015 to 2021. Spatial clusters were detected for prescription opioids, heroin, and stimulants in western Massachusetts. For synthetic opioids, hotspots became more ubiquitous across the state from 2016 to 2021, intensifying in southeastern Massachusetts. CONCLUSION: Our novel approach uncovered new time varying and spatial patterns in fatal OOD rates not previously reported. Identified shifts in fatal OOD rates by sex, age, and race/ethnicity can inform location-specific field actions targeting subpopulations at disproportionally high risk.

Generalizable biomarker prediction from cancer pathology slides with self-supervised deep learning: A retrospective multi-centric study.

Deep learning (DL) can predict microsatellite instability (MSI) from routine histopathology slides of colorectal cancer (CRC). However, it is unclear whether DL can also predict other biomarkers with high performance and whether DL predictions generalize to external patient populations. Here, we acquire CRC tissue samples from two large multi-centric studies. We systematically compare six different state-of-the-art DL architectures to predict biomarkers from pathology slides, including MSI and mutations in BRAF, KRAS, NRAS, and PIK3CA. Using a large external validation cohort to provide a realistic evaluation setting, we show that models using self-supervised, attention-based multiple-instance learning consistently outperform previous approaches while offering explainable visualizations of the indicative regions and morphologies. While the prediction of MSI and BRAF mutations reaches a clinical-grade performance, mutation prediction of PIK3CA, KRAS, and NRAS was clinically insufficient.

Analysing cerebrospinal fluid with explainable deep learning: From diagnostics to insights.

AIM: Analysis of cerebrospinal fluid (CSF) is essential for diagnostic workup of patients with neurological diseases and includes differential cell typing. The current gold standard is based on microscopic examination by specialised technicians and neuropathologists, which is time-consuming, labour-intensive and subjective. METHODS: We, therefore, developed an image analysis approach based on expert annotations of 123,181 digitised CSF objects from 78 patients corresponding to 15 clinically relevant categories and trained a multiclass convolutional neural network (CNN). RESULTS: The CNN classified the 15 categories with high accuracy (mean AUC 97.3%). By using explainable artificial intelligence (XAI), we demonstrate that the CNN identified meaningful cellular substructures in CSF cells recapitulating human pattern recognition. Based on the evaluation of 511 cells selected from 12 different CSF samples, we validated the CNN by comparing it with seven board-certified neuropathologists blinded for clinical information. Inter-rater agreement between the CNN and the ground truth was non-inferior (Krippendorff's alpha 0.79) compared with the agreement of seven human raters and the ground truth (mean Krippendorff's alpha 0.72, range 0.56-0.81). The CNN assigned the correct diagnostic label (inflammatory, haemorrhagic or neoplastic) in 10 out of 11 clinical samples, compared with 7-11 out of 11 by human raters. CONCLUSIONS: Our approach provides the basis to overcome current limitations in automated cell classification for routine diagnostics and demonstrates how a visual explanation framework can connect machine decision-making with cell properties and thus provide a novel versatile and quantitative method for investigating CSF manifestations of various neurological diseases.

Robust 6-DoF Pose Estimation under Hybrid Constraints.

To solve the problem of the insufficient accuracy and stability of the two-stage pose estimation algorithm using heatmap in the problem of occluded object pose estimation, a new robust 6-DoF pose estimation algorithm under hybrid constraints is proposed in this paper. First, a new loss function suitable for heatmap regression is formulated to improve the quality of the predicted heatmaps and increase keypoint accuracy in complex scenes. Second, the heatmap regression network is expanded and a translation regression branch is added to constrain the pose further. Finally, a robust pose optimization module is used to fuse the heatmap and translation estimates and improve the pose estimation accuracy. The proposed algorithm achieves ADD(-S) accuracy rates of 93.5% and 46.2% on the LINEMOD dataset and the Occlusion LINEMOD dataset, which are better than other state-of-the-art algorithms. Compared with the conventional two-stage heatmap-based pose estimation algorithms, the mean estimation error is greatly reduced, and the stability of pose estimation is improved. The proposed algorithm can run at a maximum speed of 22 FPS, thus constituting both a performant and efficient method.

Risk heatmaps as visual displays: Opening movie studios after the COVID-19 shutdown.

Upon shutting down operations in early 2020 due to the COVID-19 pandemic, the movie industry assembled teams of experts to help develop guidelines for returning to operation. It resulted in a joint report, The Safe Way Forward, which was created in consultation with union members and provided the basis for negotiations with the studios. A centerpiece of the report was a set of heatmaps displaying SARS-CoV-2 risks for a shoot, as a function of testing rate, community infection prevalence, community transmission rate (R0), and risk measure (either expected number of cases or probability of at least one case). We develop and demonstrate a methodology for evaluating such complex displays, in terms of how well they inform potential users, in this case, workers deciding whether the risks of a shoot are acceptable. We ask whether individuals making hypothetical return-to-work decisions can (a) read display entries, (b) compare display entries, and (c) make inferences based on display entries. Generally speaking, respondents recruited through the Amazon MTurk platform could interpret the display information accurately and make coherent decisions, suggesting that heatmaps can communicate complex risks to lay audiences. Although these heatmaps were created for practical, rather than theoretical, purposes, these results provide partial support for theoretical accounts of visual information processing and identify challenges in applying them to complex settings.

Sequential paired covariance for improved visualization of mass spectrometry imaging datasets.

Untargeted analyses in mass spectrometry imaging produce hundreds of ion images representing spatial distributions of biomolecules in biological tissues. Due to the large diversity of ions detected in untargeted analyses, normalization standards are often difficult to implement to account for pixel-to-pixel variability in imaging studies. Many normalization strategies exist to account for this variability, but they largely do not improve image quality. In this study, we present a new approach for improving image quality and visualization of tissue features by application of sequential paired covariance (SPC). This approach was demonstrated using previously published tissue datasets such as rat brain and human prostate with different biomolecules like metabolites and N-linked glycans. Data transformation by SPC improved ion images resulting in increased smoothing of biological features compared with commonly used normalization approaches.

Ethnomedicinal landscape: distribution of used medicinal plant species in Nepal.

BACKGROUND: The risk of losing traditional knowledge of medicinal plants and their use and conservation is very high. Documenting knowledge on distribution and use of medicinal plants by different ethnic groups and at spatial scale on a single platform is important from a conservation planning and management perspective. The sustainable use, continuous practice, and safeguarding of traditional knowledge are essential. Communication of such knowledge among scientists and policy makers at local and global level is equally important, as the available information at present is limited and scattered in Nepal. METHODS: In this paper, we aimed to address these shortcomings by cataloguing medicinal plants used by indigenous ethnic groups in Nepal through a systematic review of over 275 pertinent publications published between 1975 and July 2021. The review was complemented by field visits made in 21 districts. We determined the ethnomedicinal plants hotspots across the country and depicted them in heatmaps. RESULTS: The heatmaps show spatial hotspots and sites of poor ethnomedicinal plant use documentation, which is useful for evaluating the interaction of geographical and ethnobotanical variables. Mid-hills and mountainous areas of Nepal hold the highest number of medicinal plant species in use, which could be possibly associated with the presence of higher human population and diverse ethnic groups in these areas. CONCLUSION: Given the increasing concern about losing medicinal plants due to changing ecological, social, and climatic conditions, the results of this paper may be important for better understanding of how medicinal plants in use are distributed across the country and often linked to specific ethnic groups.

TSF-transformer: a time series forecasting model for exhaust gas emission using transformer.

Monitoring and prediction of exhaust gas emissions for heavy trucks is a promising way to solve environmental problems. However, the emission data acquisition is time delayed and the pattern of emission is usually irregular, which makes it very difficult to accurately predict the emission state. To deal with these problems, in this paper, we interpret emission prediction as a time series prediction problem and explore a deep learning model, a time-series forecasting Transformer (TSF-Transformer) for exhaust gas emission prediction. The exhaust emission of the heavy truck is not directly predicted, but indirectly predicted by predicting the temperature and pressure changes of the exhaust pipe under the working state of the truck. The basis of our research is based on real-time data feeds from temperature and pressure sensors installed on the exhaust pipe of approximately 12,000 heavy trucks. Therefore, the task of time series forecasting consists of two key stages: monitoring and prediction. The former utilizes the server to receive the data sent by the sensors in real-time, and the latter uses these data as samples for network training and testing. The training of the network throughout the prediction process is done in an unsupervised manner. Also, to visualize the forecast results, we weight the forecast data with the truck trajectories and present them as heatmaps. To the best of our knowledge, this is the first case of using the Transformer as the core component of the prediction model to complete the task of exhaust emissions prediction from heavy trucks. Experiments show that the prediction model outperforms other state-of-the-art methods in prediction accuracy.

Improving disease classification performance and explainability of deep learning models in radiology with heatmap generators.

As deep learning is widely used in the radiology field, the explainability of Artificial Intelligence (AI) models is becoming increasingly essential to gain clinicians' trust when using the models for diagnosis. In this research, three experiment sets were conducted with a U-Net architecture to improve the disease classification performance while enhancing the heatmaps corresponding to the model's focus through incorporating heatmap generators during training. All experiments used the dataset that contained chest radiographs, associated labels from one of the three conditions ["normal", "congestive heart failure (CHF)", and "pneumonia"], and numerical information regarding a radiologist's eye-gaze coordinates on the images. The paper that introduced this dataset developed a U-Net model, which was treated as the baseline model for this research, to show how the eye-gaze data can be used in multi-modal training for explainability improvement and disease classification. To compare the classification performances among this research's three experiment sets and the baseline model, the 95% confidence intervals (CI) of the area under the receiver operating characteristic curve (AUC) were measured. The best method achieved an AUC of 0.913 with a 95% CI of [0.860, 0.966]. "Pneumonia" and "CHF" classes, which the baseline model struggled the most to classify, had the greatest improvements, resulting in AUCs of 0.859 with a 95% CI of [0.732, 0.957] and 0.962 with a 95% CI of [0.933, 0.989], respectively. The decoder of the U-Net for the best-performing proposed method generated heatmaps that highlight the determining image parts in model classifications. These predicted heatmaps, which can be used for the explainability of the model, also improved to align well with the radiologist's eye-gaze data. Hence, this work showed that incorporating heatmap generators and eye-gaze information into training can simultaneously improve disease classification and provide explainable visuals that align well with how the radiologist viewed the chest radiographs when making diagnosis.

Ten Fast Transfer Learning Models for Carotid Ultrasound Plaque Tissue Characterization in Augmentation Framework Embedded with Heatmaps for Stroke Risk Stratification.

Background and Purpose: Only 1-2% of the internal carotid artery asymptomatic plaques are unstable as a result of >80% stenosis. Thus, unnecessary efforts can be saved if these plaques can be characterized and classified into symptomatic and asymptomatic using non-invasive B-mode ultrasound. Earlier plaque tissue characterization (PTC) methods were machine learning (ML)-based, which used hand-crafted features that yielded lower accuracy and unreliability. The proposed study shows the role of transfer learning (TL)-based deep learning models for PTC. Methods: As pertained weights were used in the supercomputer framework, we hypothesize that transfer learning (TL) provides improved performance compared with deep learning. We applied 11 kinds of artificial intelligence (AI) models, 10 of them were augmented and optimized using TL approaches-a class of Atheromatic™ 2.0 TL (AtheroPoint™, Roseville, CA, USA) that consisted of (i-ii) Visual Geometric Group-16, 19 (VGG16, 19); (iii) Inception V3 (IV3); (iv-v) DenseNet121, 169; (vi) XceptionNet; (vii) ResNet50; (viii) MobileNet; (ix) AlexNet; (x) SqueezeNet; and one DL-based (xi) SuriNet-derived from UNet. We benchmark 11 AI models against our earlier deep convolutional neural network (DCNN) model. Results: The best performing TL was MobileNet, with accuracy and area-under-the-curve (AUC) pairs of 96.10 ± 3% and 0.961 (p < 0.0001), respectively. In DL, DCNN was comparable to SuriNet, with an accuracy of 95.66% and 92.7 ± 5.66%, and an AUC of 0.956 (p < 0.0001) and 0.927 (p < 0.0001), respectively. We validated the performance of the AI architectures with established biomarkers such as greyscale median (GSM), fractal dimension (FD), higher-order spectra (HOS), and visual heatmaps. We benchmarked against previously developed Atheromatic™ 1.0 ML and showed an improvement of 12.9%. Conclusions: TL is a powerful AI tool for PTC into symptomatic and asymptomatic plaques.

Two decades of advances in muscle imaging in children: from pattern recognition of muscle diseases to quantification and machine learning approaches.

Muscle imaging has progressively gained popularity in the neuromuscular field. Together with detailed clinical examination and muscle biopsy, it has become one of the main tools for deep phenotyping and orientation of etiological diagnosis. Even in the current era of powerful new generation sequencing, muscle MRI has arisen as a tool for prioritization of certain genetic entities, supporting the pathogenicity of variants of unknown significance and facilitating diagnosis in cases with an initially inconclusive genetic study. Although the utility of muscle imaging is increasingly clear, it has not reached its full potential in clinical practice. Pattern recognition is known for a number of diseases and will certainly be enhanced by the use of machine learning approaches. For instance, MRI heatmap representations might be confronted with molecular results by obtaining a probabilistic diagnosis based in each disease "MRI fingerprints". Muscle ultrasound as a screening tool and quantified techniques such as Dixon MRI seem still underdeveloped. In this paper, we aim to appraise the advances in recent years in pediatric muscle imaging and try to define areas of uncertainty and potential advances that might become standardized to be widely used in the future.

Spatial Bias in Antibody Microarrays May Be an Underappreciated Source of Variability.

Antibody microarrays enable multiplexed protein detection with minimal reagent consumption, but they continue to be plagued by lack of reproducibility. Chemically functionalized glass slides are used as substrates, yet antibody binding spatial inhomogeneity across the slide has not been analyzed in antibody microarrays. Here, we characterize spatial bias across five commercial slides patterned with nine overlapping dense arrays (by combining three buffers and three different antibodies), and we measure signal variation for both antibody immobilization and the assay signal, generating 270 heatmaps. Spatial bias varied across models, and the coefficient of variation ranged from 4.6 to 50%, which was unexpectedly large. Next, we evaluated three layouts of spot replicates-local, random, and structured random-for their capacity to predict assay variation. Local replicates are widely used but systematically underestimate the whole-slide variation by up to seven times; structured random replicates gave the most accurate estimation. Our results highlight the risk and consequences of using local replicates: the underappreciation of spatial bias as a source of variability, poor assay reproducibility, and possible overconfidence in assay results. We recommend the detailed characterization of spatial bias for antibody microarrays and the description and use of distributed positive replicates for research and clinical applications.

The design heatmap: A simple visualization of D -optimality design problems.

Optimal experimental designs are often formal and specific, and not intuitively plausible to practical experimenters. However, even in theory, there often are many different possible design points providing identical or nearly identical information compared to the design points of a strictly optimal design. In practical applications, this can be used to find designs that are a compromise between mathematical optimality and practical requirements, including preferences of experimenters. For this purpose, we propose a derivative-based two-dimensional graphical representation of the design space that, given any optimal design is already known, will show which areas of the design space are relevant for good designs and how these areas relate to each other. While existing equivalence theorems already allow such an illustration in regard to the relevance of design points only, our approach also shows whether different design points contribute the same kind of information, and thus allows tweaking of designs for practical applications, especially in regard to the splitting and combining of design points. We demonstrate the approach on a toxicological trial where a D -optimal design for a dose-response experiment modeled by a four-parameter log-logistic function was requested. As these designs require a prior estimate of the relevant parameters, which is difficult to obtain in a practical situation, we also discuss an adaption of our representations to the criterion of Bayesian D -optimality. While we focus on D -optimality, the approach is in principle applicable to different optimality criteria as well. However, much of the computational and graphical simplicity will be lost.