Incorrect Assignment Research Articles

Comparison of Illumina and Oxford Nanopore sequencing data quality for Clostridioides difficile genome analysis and their application for epidemiological surveillance

BackgroundThe burden of Clostridioides difficile as a nosocomial- and community-acquired pathogen has been increasing over the recent decades, including reports of severe outbreaks. Molecular and virulence genotyping are central for the epidemiological surveillance of this pathogen, but need to balance accuracy and rapid turnaround time of the results. While Illumina short-read sequencing has been adopted as the gold standard to investigate C. difficile virulence and transmission routes, little is known about the potential of Nanopore long-read sequencing in this field. The goal of our study was to compare sequencing and assembly quality of 37 C. difficile isolates using Illumina (SPAdes assembled) and Nanopore (Flye and Unicycler assembled) data alone, along with hybrid assemblies obtained with short-read polishing of long reads.ResultsIllumina sequencing produced reads with an average quality of 99.68% (Q25), while Nanopore sequencing produced reads reaching an average quality of 96.84% (Q15), showing a tenfold difference in quality. Sequence type (ST) designation from Nanopore assemblies failed to detect ST5, ST7, ST8, ST13 and ST49, while ST designation based on unpolished Nanopore reads using Krocus was successful for all STs. Nanopore sequences exhibited an average of 640 base errors per genome (~ 0.015% substitution rate), which was reflected by the incorrect assignment of over 180 alleles in core genome multilocus sequence typing (cgMLST) analysis. As a result, Nanopore-derived phylogenies were not as accurate as the Illumina reference, and therefore inadequate for precise investigation of transmission events. Both sequencing platforms provided comparable, satisfactory results for the detection of virulence genes tcdA, tcdB, cdtAB and in-frame deletions in tcdC.ConclusionCompared to Illumina, Nanopore has higher error rate, which limits its application for high-resolution epidemiological surveillance. However, the short analysis time, lower cost and more simple procedure combined with correctly identified STs and virulence genes, makes it an alternative when fast and less detailed analyses are preferred.

Protonation effects in protein-ligand complexes - a case study of endothiapepsin and pepstatin A with computational and experimental methods.

Protonation states serve as an essential molecular recognition motif for biological processes. Their correct consideration is key to successful drug design campaigns, since chemoinformatic tools usually deal with default protonation states of ligands and proteins and miss atypical protonation states. The protonation pattern for the Endothiapepsin/PepstatinA (EP/pepA) complex is investigated using different dry lab and wet lab techniques. ITC experiments revealed an uptake of more than one mole of protons upon pepA binding to EP. Since these experiments were performed at physiological conditions (and not at pH=4 at which a large variety of crystal structures is available), a novel crystal structure at pH=7.6 was determined. This crystal structure showed that only modest structural changes occur upon increasing the pH value. This lead to computational studies (Poisson Boltzmann calculations and constant pH MD simulations)to reveal the exact location of the protonation event. Both computational studies could reveal a significant pKa shift resulting in non-default protonation state and that the catalytic dyad is responsible for the uptake of protons. This study shows that assessing protonation states for two separate systems (protein and ligand) might result in the incorrect assignment of protonation states and hence incorrect calculation of binding energy.

Antenatal Diagnosis and Treatment in Congenital Adrenal Hyperplasia Due to 21-hydroxylase Deficiency and Congenital Adrenal Hyperplasia Screening in Newborns

Signs of virilization, such as clitoromegaly, labio-scrotal fusion, and urogenital sinus may be observed in females with 21-hydroxylase deficiency (21-OHD) and other rare virilizing forms of congenital adrenal hyperplasia (CAH). This makes sex determination difficult, and multiple reconstructive surgeries in the postnatal period may be required. As 21-OHD is an autosomal recessive disease, the chance of any child being affected is one in four and so only one in eight will be an affected female. The primary objective of antenatal diagnosis is to identify only the affected fetus in the early gestational weeks before the onset of genital organogenesis and to treat that case. Therefore, studies aimed at antenatal diagnosis and preventing adrenal androgen exposure in the female fetus with CAH have long been of interest. Antenatal steroid treatment is considered experimental and controversial for safety reasons in recent clinical guidelines. If antenatal treatment is to be used, it is recommended that it should be performed in experienced centers that can collect data on a large number of cases which will help to define the benefits and harms of treatment better. In the postnatal period, a severe deficiency of the 21-hydroxylase enzyme leads to life-threatening adrenocortical insufficiency in both sexes and varying degrees of pathology of the external genitalia in females. This condition is also associated with high mortality in the first days of life and an increased risk of incorrect sex assignment. Neonatal screening for 21-OHD CAH effectively detects the severe forms and reduces mortality, and it is instrumental in the correct sex assignment of female cases.

Alevin-fry-atac enables rapid and memory frugal mapping of single-cell ATAC-seq data using virtual colors for accurate genomic pseudoalignment.

Ultrafast mapping of short reads to transcriptomic and metagenomic references via lightweight mapping techniques such as pseudoalignment has demonstrated success in substantially accelerating several types of analyses without much loss in accuracy compared to alignment-based approaches. The application of pseudoalignment to large reference sequences - like the genome - is, however, not trivial, due to the large size of the references or "targets" (i.e. chromosomes) and the presence of repetitive sequences within an individual reference sequence. This can lead to multiple matching locations for a -mer within a single reference, which in turn can lead to false positive mappings and incorrect reference assignments for a read when the colors across the -mer matches for a read are aggregated. Even when the read is determined to map to the appropriate reference, the increased occurrence of -mer multi-matches within a reference can prevent the determination of the correct approximate position of the read, which is often critical in applications that map short reads to the genome.

Clinical Applicability of 2-Field High-Resolution and Extended HLA-Allele Typing in Deceased Donor Kidney Allocation.

HLA-compatibility remains an important triage test for deceased donor kidney allocation. Low-intermediate resolution donor HLA-typing is typically available at allocation, but its accuracy in assigning pre-transplant donor-specific anti-HLA antibody (DSA) and HLA mismatches compared to 2-field high-resolution typing is poorly characterised. Consecutive deceased donor/recipient pairs from a single centre between 2016 and 2020 were included. Majority of donor typing at HLA-ABDRB1 loci were performed at low-intermediate resolution, with 2-field high-resolution NGS typing across extended loci performed by NGS-technique post-transplantation. We compared the two typing methods for (1) accuracy of pre-transplant DSA assignment; (2) misassignment of HLA-antigen/allele mismatches and performance of each model for acute rejection and (3) proportion of recipients who developed de novo DSA (dnDSA) when matched at antigen but mismatched at allele level. Of 179 deceased donor/recipient pairs, 157 donors had low-intermediate resolution typing and 22 with high-resolution ONT typing. Sixty-two recipients (35%) had potential pre-transplant DSAs, with incorrect assignment of allele-specific Class I and II actual DSAs in 31% and 53% of cases, respectively. NGS typing identified 59 (33%) additional HLA-DRB1 allele mismatches. ONT typing accurately assigned pre-transplant DSAs and allele mismatches in all cases. Seven (4%) recipients with antigen/allele level discordance developed dnDSAs, majority HLA-DQ antibodies. Two-field high-resolution donor HLA typing may provide a more accurate transplant immunological risk assessment and identify those at risk of developing dnDSA to matched HLA antigen.

Performance degradation assessment of rolling bearings for electrostatic monitoring based on IDDAE and ADPC

Abstract To improve the early degradation detection capability of the electrostatic monitoring system for rolling bearings, a performance degradation evaluation method based on improved deep denoising autoencoder and adaptive density peak clustering (ADPC) is proposed in this paper. Firstly, the fusion of electrostatic charge signal features with conventional time-domain, frequency-domain and time–frequency-domain features constitutes the characteristic parameter set of the electrostatic monitoring system indicating the status of the bearings. Then, in order to improve the feature extraction ability of DAE, the deep network DDAE is constructed, and L1 regularisation and Dropout mechanism are applied to avoid overfitting in the deep network, so as to achieve non-linear mapping dimensionality reduction of high-dimensional features. Moreover, to eliminate the error caused by manually selecting the clustering centre, the parameters are adaptively determined by entropy value method and comprehensive optimisation search on the basis of DPC, thus avoiding the ‘chain effect’ that occurs in traditional DPC when data are incorrectly aggregated due to incorrect assignment of clustering centres. Consequently, an improved ADPC algorithm is used to establish a model to measure the health status of bearings, calculate the Mahalanobis distance (MD) between the test set and the cluster centre, and quantitatively characterize the degree of performance degradation of rolling bearings. Finally, combining the 3 δ principle, a repair method that can satisfy online monitoring and adapt to spurious fluctuations in different situations is established on a sliding window to obtain an improved index IMD that can accurately characterise the rolling bearing degradation process. The experimental results show that the proposed method can identify early bearing degradation earlier and has better monotonicity, robustness and tendency than other performance degradation assessment methods.

Exploring the relationship between functional diversity and growth in VC-backed new technology ventures

PurposeDespite significant research attention to top management team diversity, the potential influence of other employees’ diversity on venture growth at different maturity stages such as middle management teams and nonmanagerial employees at the base of the organizational hierarchy remains largely overlooked. This study explores this relationship in new technology venture development, with a focus on the influence of functional skills diversity across different organizational levels and maturity stages.Design/methodology/approachThis study uses a linked employer–employee dataset covering new technology ventures in the digital industry that spans from 2010 to 2020. The sample includes 296 VC-backed new ventures located in the greater Paris (France) area. Through a problem-solving lens, the study examines functional skills at various organizational levels, including 5,243 top managers, 10,274 middle managers and 29,306 nonmanagerial employees.FindingsResults indicate that placing exclusive emphasis on top-level managers could lead to incorrect assignment of diversity effects as such effects are likely shared with lower organizational levels. In addition, the findings demonstrate that the diversity–venture growth relationship varies across new ventures’ funding stages, with a notably stronger correlation in the late stages of financing.Practical implicationsThis study demonstrates that promoting functional diversity across all organizational levels, not just top management, may boost firms’ growth. Furthermore, the effectiveness of functional diversity varies depending on the firm’s financing stage.Originality/valueThis study constructs a multilayered organizational framework to explore the relationship between diversity and new technology venture growth, using a fully reproducible skills diversity classification methodology that is applicable for future investigations.

Misdiagnosis and underdiagnosis of multiple sclerosis: A systematic review and meta-analysis

Background: Diagnostic errors in multiple sclerosis (MS) impact patients and healthcare systems. Objectives: This study aimed to determine the prevalence of MS misdiagnosis and underdiagnosis, time delay in reaching a correct diagnosis and potential impact of sex. Methods: Systematic review and meta-analysis on MS diagnostic errors. Results: Out of 3910 studies, we included 62 for a qualitative synthesis and 24 for meta-analyses. Frequency of misdiagnosis (incorrect assignment of an MS diagnosis) ranged from 5% to 41%, with a pooled proportion based on six studies of 15% (95% CI: 9%–26%, n = 1621). The delay to rectify a misdiagnosis ranged from 0.3 to 15.9 years. Conversely, underdiagnosis (unrecognized diagnosis of MS) ranged from 3% to 58%, with a pooled proportion in four studies of 36% (95% CI: 20%–55%, n = 728). Pooling seven studies comprising 2851 individuals suggested a diagnostic delay to establish a correct MS diagnosis of 17.3 months (95% CI: 11.9–22.7) in patients underdiagnosed. In a meta-analysis of five studies, women were 2.1 times more likely to be misdiagnosed with MS compared to men (odds ratio, 95% CI: 1.53–2.86). Conclusion: This study provides summary-level evidence for the high prevalence of MS misdiagnosis and underdiagnosis. Future studies are needed to understand the causes of these diagnostic challenges in MS care.

Artificial neural networks with better analysis reliability in data mining

If there are relatively few cases, semi-supervised learning approaches make advantage of a large amount of unlabeled data to assist develop a better classifier. To expand the labeled training set and update the classifier, a fundamental method is to select and label the unlabeled instances for which the current classifier has higher classification confidence. This approach is primarily used in two distinct semi-supervised learning paradigms: co-training and self-training. However, compared to self-labeled examples that would be tagged by a classifier, the real labeled instances will be more trustworthy. Incorrect label assignment to unlabeled occurrences might potentially compromise the classifier's accuracy in classification. This research presents a novel instance selection method based on actual labeled data. This will take into account the classifier's current performance on unlabeled data in addition to its performance on actual labeled data alone. This uses the accuracy changes in the newly trained classifier over the original labeled data as a criterion in each iteration to determine whether or not the selected most confident unlabeled examples would be accepted by a subsequent iteration. Naïve Bayes (NB) will be used as the basic classifier in the co-training and self-training studies. The findings indicate that the accuracy and categorization of self-training and co-training will be greatly enhanced by SIS. As compared to semi-supervised classification methods, it will enhance accuracy, precision, recall, and F1 score, according to the findings.

Fluorescent EEM-PARAFAC considerations of Aldrich humic acid salt as an aquatic or terrestrial organic matter surrogate

The fluorescence of “long dead” terrestrially sourced organic carbon (Aldrich humic acid salt) was compared to “recently dead or excreted” aquatically sourced organic carbon (ornamental pond). This research eliminated the “living” (biologically active) category of organic carbon (by azide treatment and filtration) which prevented organic matter continuing to metabolize due to microbial activity during storage and analysis. The focus in this context is on the diagenetic (fulvic and humic) substances, which are considered the most recalcitrant organic carbon forms in water treatment. The purposes of this research were to investigate (a) what excitation-emission matrix (EEM) spectroscopy and PARAllel FACtor (PARAFAC) analysis (EEM-PARAFAC) modeling reveals about the limitations to the use of terrestrially derived Aldrich Humic Acid as an aquatic organic matter surrogate, (b) potential flaws observed in EEM spectra and incorrect PARAFAC component assignments gathered in contemporaneous studies, and (c) how EEM-PARAFAC modeling can be used to illustrate the continuum in organic matter aging or maturity. Additionally, to expand the analysis, data were compared to previously published PARAFAC models derived from other freshwater sources. An important result of this research was discovery of a PARAFAC component peak observed at the emission (EM) wavelength at 540 nm in Aldrich Humic Acid which has been rarely noted in the literature, and not previously recognized for its potential importance. This research implies that Aldrich humic acid should only be used as an aquatic surrogate with caution, and that it can be a surrogate for terrestrial soil or ore samples.

When is a reported air-stable PrIV complex not a PrIV complex? Answer: when it is PrIII

All the data recently reported for the purported air/water stable PrIV complex, [(L)PrIV(L)4(OH0.94Cl0.06)2PrIV(L)]n (L = N-acetylanthranilate) [1], have been reinterpreted to show that it is, in fact, the PrIII complex [Pr2L6(H2O)2]n. Importantly, it is isomorphous with its reported EuIII analogue [Eu2L6(H2O)2]n [2], providing further evidence of the incorrect assignment in the literature.

"Covering provider": an effort to streamline clinical communication chaos.

This report describes a root cause analysis of incorrect provider assignments and a standardized workflow developed to improve the clarity and accuracy of provider assignments. A multidisciplinary working group involving housestaff was assembled. Key drivers were identified using value stream mapping and fishbone analysis. A report was developed to allow for the analysis of correct provider assignments. A standardized workflow was created and piloted with a single service line. Pre- and post-pilot surveys were administered to nursing staff and participating housestaff on the unit. Four key drivers were identified. A standardized workflow was created with an exclusive treatment team role in Epic held by a single provider at any given time, with a corresponding patient list column displaying provider information for each patient. Pre- and post-survey responses report decreased confusion, decreased provider identification errors, and increased user satisfaction among RNs and residents with sustained uptake over time. This work demonstrates structured root cause analysis, notably engaging housestaff, to develop a standardized workflow for an understudied and growing problem. The development of tools and strategies to address the widespread burdens resulting from clinical communication failures is needed.

Validity of ICD-10 diagnosis codes for placenta accreta spectrum disorders.

The 10th revision of the International Classification of Diseases, Clinical Modification (ICD-10) includes diagnosis codes for placenta accreta spectrum for the first time. These codes could enable valuable research and surveillance of placenta accreta spectrum, a life-threatening pregnancy complication that is increasing in incidence. We sought to evaluate the validity of placenta accreta spectrum diagnosis codes that were introduced in ICD-10 and assess contributing factors to incorrect code assignments. We calculated sensitivity, specificity, positive predictive value and negative predictive value of the ICD-10 placenta accreta spectrum code assignments after reviewing medical records from October 2015 to March 2020 at a quaternary obstetric centre. Histopathologic diagnosis was considered the gold standard. Among 22,345 patients, 104 (0.46%) had an ICD-10 code for placenta accreta spectrum and 51 (0.23%) had a histopathologic diagnosis. ICD-10 codes had a sensitivity of 0.71 (95% CI 0.56, 0.83), specificity of 0.98 (95% CI 0.93, 1.00), positive predictive value of 0.61 (95% CI 0.48, 0.72) and negative predictive value of 1.00 (95% CI 0.96, 1.00). The sensitivities of the ICD-10 codes for placenta accreta spectrum subtypes- accreta, increta and percreta-were 0.55 (95% CI 0.31, 0.78), 0.33 (95% CI 0.12, 0.62) and 0.56 (95% CI 0.31, 0.78), respectively. Cases with incorrect code assignment were less morbid than cases with correct code assignment, with a lower incidence of hysterectomy at delivery (17% vs 100%), blood transfusion (26% vs 75%) and admission to the intensive care unit (0% vs 53%). Primary reasons for code misassignment included code assigned to cases of occult placenta accreta (35%) or to cases with clinical evidence of placental adherence without histopatholic diagnostic (35%) features. These findings from a quaternary obstetric centre suggest that ICD-10 codes may be useful for research and surveillance of placenta accreta spectrum, but researchers should be aware of likely substantial false positive cases.

Development of the multivariate administrative data cystectomy model and its impact on misclassification bias

BackgroundMisclassification bias (MB) is the deviation of measured from true values due to incorrect case assignment. This study compared MB when cystectomy status was determined using administrative database codes vs. predicted cystectomy probability.MethodsWe identified every primary cystectomy-diversion type at a single hospital 2009–2019. We linked to claims data to measure true association of cystectomy with 30 patient and hospitalization factors. Associations were also measured when cystectomy status was assigned using billing codes and by cystectomy probability from multivariate logistic regression model with covariates from administrative data. MB was the difference between measured and true associations.Results500 people underwent cystectomy (0.12% of 428 677 hospitalizations). Sensitivity and positive predictive values for cystectomy codes were 97.1% and 58.6% for incontinent diversions and 100.0% and 48.4% for continent diversions, respectively. The model accurately predicted cystectomy-incontinent diversion (c-statistic [C] 0.999, Integrated Calibration Index [ICI] 0.000) and cystectomy-continent diversion (C:1.000, ICI 0.000) probabilities. MB was significantly lower when model-based predictions was used to impute cystectomy-diversion type status using for both incontinent cystectomy (F = 12.75; p < .0001) and continent cystectomy (F = 11.25; p < .0001).ConclusionsA model using administrative data accurately returned the probability that cystectomy by diversion type occurred during a hospitalization. Using this model to impute cystectomy status minimized MB. Accuracy of administrative database research can be increased by using probabilistic imputation to determine case status instead of individual codes.

Expression of concern: Natural copy number variation of tandemly repeated regulatory SNORD RNAs leads to individual phenotypic differences in mice.

Keshavarz, M., Savriama, Y., Refki, P., Reeves, R. G., & Tautz, D. (2021). Natural copy number variation of tandemly repeated regulatory SNORD RNAs leads to individual phenotypic differences in mice. Molecular Ecology, 30, 4708-4722. https://doi.org/10.1111/mec.16076 This expression of concern is for the above article, published online on 12 July 2021 in Wiley Online Library (wileyonlinelibrary.com), and has been published by agreement between the journal Editor-in-Chief, Loren Rieseberg and John Wiley & Sons Ltd. The expression of concern follows an internal investigation by the Max-Planck Society, which found a number of inconsistencies in the supporting information of the article, later also confirmed by the authors: Incorrect sex assignment for a set of mouse IDs (reported in TableS1, SNORD copy numbers) Inconsistencies in the ddPCR data that were used to determine SNORD copy number variation (reported in Tables S1, S4 and S6): Over-representation of certain values for SD, incorrect calculation of averages and SD values (TableS1, SNORD copy numbers). SNORD115 and SNORD116 copy numbers differed between columns C and I and columns D and J, respectively, despite having been determined only once (TableS4, other CNVs). Incorrect mouse IDs for family VI and VIII (TableS6, family study). Replicate three for SNORD115 and SNORD116 determination is often identical to the average (TableS6, family study). D.T. and M.K. could not retrieve the relevant raw data for the published ddPCR experiments that would have allowed to resolve the inconsistencies based on a re-analysis. Consequently, it cannot be excluded that these irregularities affect the reproducibility of the study and the conclusions that are based on these data. Therefore, the journal has decided to issue an expression of concern to inform and alert the readers. The transcriptome analyses and morphometric data presented in the article remain valid, including the corresponding original data depositions at ENA (accession number PRJEB29473) and Zenodo (record 4750371). The issues with the incorrect sex assignment (TableS1) and the erroneous mouse IDs (TableS6) were resolved, the corrected Tables S1 and S6 are provided in the updated Supporting Information.

A novel incomplete hesitant fuzzy information supplement and clustering method for large-scale group decision-making.

Clustering is an effective means to reduce the scaling of large-scale group decision-making (LSGDM). However, there are many problems with clustering methods, such as incomplete or ambiguous information usually provided by different decision makers. Traditional clustering methods may not be able to handle these situations effectively, resulting in incomplete decision-making information. Calculating the clustering centers may become very complex and time-consuming. Inappropriate distance weights may also lead to incorrect cluster assignments, and these problems will seriously affect the clustering results. This research provides a novel incomplete hesitant fuzzy information supplement and clustering approach for large-scale group decision-making in order to address the aforementioned difficulties. First, the approach takes into account the trust degradation and the inhibition of relationships of distrust in the process of trust propagation, and then it builds a global and local network of trust. A novel supplemental formula is provided that takes into account the decision-preference maker's as well as the trust-neighbor's information, allowing the decision-neighbor maker's recommendation to be realized. Therefore, an improved distance function can be proposed to calculate the weights by combining the relative standard deviation theory and selecting the selected clustering centers by using the density peaks in order to optimize the selection of clustering centers and reduce the complexity and scaling of the decision. Finally, an example is presented to demonstrate how the proposed method can be applied. The consistency index and comparison experiments are used to evaluate if the suggested approach is effective and reliable.

A cautionary tale of low-pass sequencing and imputation with respect to haplotype accuracy

BackgroundLow-pass whole-genome sequencing and imputation offer significant cost savings, enabling substantial increases in sample size and statistical power. This approach is particularly promising in livestock breeding, providing an affordable means of screening individuals for deleterious alleles or calculating genomic breeding values. Consequently, it may also be of value in companion animal genomics to support pedigree breeding. We sought to evaluate in dogs the impact of low coverage sequencing and reference-guided imputation on genotype concordance and association analyses.ResultsDNA isolated from saliva of 30 Labrador retrievers was sequenced at low (0.9X and 3.8X) and high (43.5X) coverage, and down-sampled from 43.5X to 9.6X and 17.4X. Genotype imputation was performed using a diverse reference panel (1021 dogs), and two subsets of the former panel (256 dogs each) where one had an excess of Labrador retrievers relative to other breeds. We observed little difference in imputed genotype concordance between reference panels. Association analyses for a locus acting as a disease proxy were performed using single-marker (GEMMA) and haplotype-based (XP-EHH) tests. GEMMA results were highly correlated (r ≥ 0.97) between 43.5X and ≥ 3.8X depths of coverage, while for 0.9X the correlation was lower (r ≤ 0.8). XP-EHH results were less well correlated, with r ranging from 0.58 (0.9X) to 0.88 (17.4X). Across a random sample of 10,000 genomic regions averaging 17 kb in size, we observed a median of three haplotypes per dog across the sequencing depths, with 5% of the regions returning more than eight haplotypes. Inspection of one such region revealed genotype and phasing inconsistencies across sequencing depths.ConclusionsWe demonstrate that saliva-derived canine DNA is suitable for whole-genome sequencing, highlighting the feasibility of client-based sampling. Low-pass sequencing and imputation require caution as incorrect allele assignments result when the subject possesses alleles that are absent in the reference panel. Larger panels have the capacity for greater allelic diversity, which should reduce the potential for imputation error. Although low-pass sequencing can accurately impute allele dosage, we highlight issues with phasing accuracy that impact haplotype-based analyses. Consequently, if accurately phased genotypes are required for analyses, we advocate sequencing at high depth (> 20X).

Prompt-driven Latent Domain Generalization for Medical Image Classification.

Deep learning models for medical image analysis easily suffer from distribution shifts caused by dataset artifact bias, camera variations, differences in the imaging station, etc., leading to unreliable diagnoses in real-world clinical settings. Domain generalization (DG) methods, which aim to train models on multiple domains to perform well on unseen domains, offer a promising direction to solve the problem. However, existing DG methods assume domain labels of each image are available and accurate, which is typically feasible for only a limited number of medical datasets. To address these challenges, we propose a unified DG framework for medical image classification without relying on domain labels, called Prompt-driven Latent Domain Generalization (PLDG). PLDG consists of unsupervised domain discovery and prompt learning. This framework first discovers pseudo domain labels by clustering the bias-associated style features, then leverages collaborative domain prompts to guide a Vision Transformer to learn knowledge from discovered diverse domains. To facilitate cross-domain knowledge learning between different prompts, we introduce a domain prompt generator that enables knowledge sharing between domain prompts and a shared prompt. A domain mixup strategy is additionally employed for more flexible decision margins and mitigates the risk of incorrect domain assignments. Extensive experiments on three medical image classification tasks and one debiasing task demonstrate that our method can achieve comparable or even superior performance than conventional DG algorithms without relying on domain labels. Our code is publicly available at https://github.com/SiyuanYan1/PLDG/tree/main.

Is first- or third-party audience data more effective for reaching the ‘right’ customers? The case of IT decision-makers

AbstractOften marketers face the challenge of how to communicate best with the customers who have the right responsibilities, influence or purchasing power, especially in business-to-business (B2B) settings. For example, B2B marketers selling software and IT need to identify IT decision-makers (ITDMs) within organizations. The modern digital environment in theory allows marketers to target individuals in organizations through specifically designed third-party audience segments based on deterministic prospect lists or probabilistic inference. However, in this paper we show that in our context, such ‘off-the-shelf’ segments perform no better at reaching the right person than random prospecting. We present evidence that even deterministic attribute information is flawed for ITDM identification, and that the poor campaign results can be partly linked to incorrect assignment of established prospect profiles to online identifiers. We then use access to our publisher network data to investigate what would happen if the advertiser had used first-party data that are less susceptible to the identified issues. We demonstrate that first-party demographics or contextual information allows advertisers and publishers to outperform both third-party ITDM audience segments and random prospecting. Our findings have implications for understanding the shift in digital advertising away from third-party cookie tracking, and how to execute digital marketing in the context of broad privacy regulation.

‘To be’ or not ‘to be’: an analysis of copula production and omission in people with non-fluent aphasia

ABSTRACT Background Agrammatic aphasia has been widely associated with impairments with functional words and complex sentences. Speech errors of people with aphasia (PWA) have been reported to be selective, with patterns of omissions in functional words, most notably in the domain of tense inflection on verbs compared to agreement in morphologically rich languages. Aims In languages like English, where tense and agreement are hard to disentangle in their inflectional paradigms, investigations of the inflectional domain in PWA are rare. In this study, we introduce a novel approach that allows the disentangling of inflectional errors in English through the patterns of copula omission of the verb to be. The inflectional system of the functional verb to be is richer, and its distribution in the sentence is based on the semantics of its predicate (stage-level vs. individual-level). Methods and procedures Spontaneous productions of 16 PWA collated from AphasiaBank transcripts were analysed for violations of tense alongside other patterns of error that could suggest an impairment in the inflectional domain. Outcomes and results Copula deletion was found to be modulated by the semantics of the predicate, showing a selective pattern of omission in stage-level predicates. Incorrect case assignment (accusative in place of nominative) was also observed as an indicator of impaired tense. Conclusions The results confirm the effectiveness of copula to be to investigate the English inflectional system and substantiate previous studies on selective errors in the verbal domain in PWA in English.