Labeling Errors Research Articles

Comparison of methods for the detection of outliers and associated biomarkers in mislabeled omics data

BackgroundPrevious studies have reported that labeling errors are not uncommon in omics data. Potential outliers may severely undermine the correct classification of patients and the identification of reliable biomarkers for a particular disease. Three methods have been proposed to address the problem: sparse label-noise-robust logistic regression (Rlogreg), robust elastic net based on the least trimmed square (enetLTS), and Ensemble. Ensemble is an ensembled classification based on distinct feature selection and modeling strategies. The accuracy of biomarker selection and outlier detection of these methods needs to be evaluated and compared so that the appropriate method can be chosen.ResultsThe accuracy of variable selection, outlier identification, and prediction of three methods (Ensemble, enetLTS, Rlogreg) were compared for simulated and an RNA-seq dataset. On simulated datasets, Ensemble had the highest variable selection accuracy, as measured by a comprehensive index, and lowest false discovery rate among the three methods. When the sample size was large and the proportion of outliers was ≤5%, the positive selection rate of Ensemble was similar to that of enetLTS. However, when the proportion of outliers was 10% or 15%, Ensemble missed some variables that affected the response variables.Overall, enetLTS had the best outlier detection accuracy with false positive rates < 0.05 and high sensitivity, and enetLTS still performed well when the proportion of outliers was relatively large. With 1% or 2% outliers, Ensemble showed high outlier detection accuracy, but with higher proportions of outliers Ensemble missed many mislabeled samples. Rlogreg and Ensemble were less accurate in identifying outliers than enetLTS. The prediction accuracy of enetLTS was better than that of Rlogreg. Running Ensemble on a subset of data after removing the outliers identified by enetLTS improved the variable selection accuracy of Ensemble.ConclusionsWhen the proportion of outliers is ≤5%, Ensemble can be used for variable selection. When the proportion of outliers is > 5%, Ensemble can be used for variable selection on a subset after removing outliers identified by enetLTS. For outlier identification, enetLTS is the recommended method. In practice, the proportion of outliers can be estimated according to the inaccuracy of the diagnostic methods used.

Hypersensitive H2 sensor based on polymer planar Bragg gratings coated with Pt-loaded WO3-SiO2: erratum.

We present an erratum to our Letter [Opt. Lett.45, 3601 (2020)OPLEDP0146-959210.1364/OL.395341]. Labeling errors in two figures and an incorrect sentence are revised. The corrections have no influence on the conclusions of the original Letter.

Questions Can Answer Questions About Mechanisms of Preschoolers' Selective Word Learning.

This study examined how inferences about epistemic competence and generalized labeling errors influence children's selective word learning. Three- to 4-year-olds (N=128) learned words from informants who asked questions about objects, mentioning either correct or incorrect labels. Such questions do not convey stark differences in informants' epistemic competence. Inaccurate labels, however, generate error signals that can lead to weaker encoding of novel information. Preschoolers retained novel labels from both informants but were slower to respond in the Inaccurate Labeler condition. When the test procedure was not sensitive to the strength of information encoding, children performed above chance in both conditions and their response times did not differ. These results suggest that epistemic-level inferences and error generalizations influence preschoolers' selective word learning concurrently.

Application of DNA barcoding in fish identification of supermarkets in Henan province, China: More and longer COI gene sequences were obtained by designing new primers

In recent years, DNA barcode technology has been widely used in food identification, especially in the identification of fish. In China, there are few studies on the authenticity of fish products in Henan province of China. In this study, 179 fish samples were collected from supermarkets in Zhengzhou city and Xinxiang city in Henan province, China. COI gene sequences were obtained with PCR technology by designing specific primers and universal primers. COI gene sequences of all samples were obtained to identify species, which is used to investigate species substitution and mislabeling of the fish sold in the two regional markets. The molecular identification results showed that 28.49% (51/179) fish samples were not consistent with the labels. Substitution of high-price fish by low-price fish was prevalent. For example, halibut (Pleuronectiformes) and cod (Gadus) are replaced by striped catfish (Pangasianodon hypophthalmus), and some merchants label bighead carp (Hypophthalmichthys nobilis) as cod (Gadus), there are also accidental labeling errors (such as labels for greenfin horse-faced filefish (Thamnaconus septentrionalis) have been identified as grass carp (Ctenopharyngodon idella) and Wuchang bream (Megalobrama amblycephala) etc. Most of the samples labeled correctly are the fish of low economic value and the fresh fish. This study shows that almost all the commercial fish can be identified by COI DNA barcoding by newly designed primers. Finally, this study also gives a reference of real species of fish fillet in Henan province in China.

Analysis of the role and robustness of artificial intelligence in commodity image recognition under deep learning neural network.

In order to explore the application of the image recognition model based on multi-stage convolutional neural network (MS-CNN) in the deep learning neural network in the intelligent recognition of commodity images and the recognition performance of the method, in the study, the features of color, shape, and texture of commodity images are first analyzed, and the basic structure of deep convolutional neural network (CNN) model is analyzed. Then, 50,000 pictures containing different commodities are constructed to verify the recognition effect of the model. Finally, the MS-CNN model is taken as the research object for improvement to explore the influence of label errors (p = 0.03, 0.05, 0.07, 0.09, 0.12) with different parameter settings and different probabilities (size of convolutional kernel, Dropout rate) on the recognition accuracy of MS-CNN model, at the same time, a CIR system platform based on MS-CNN model is built, and the recognition performance of salt and pepper noise images with different SNR (0, 0.03, 0.05, 0.07, 0.1) was compared, then the performance of the algorithm in the actual image recognition test was compared. The results show that the recognition accuracy is the highest (97.8%) when the convolution kernel size in the MS-CNN model is 2*2 and 3*3, and the average recognition accuracy is the highest (97.8%) when the dropout rate is 0.1; when the error probability of picture label is 12%, the recognition accuracy of the model constructed in this study is above 96%. Finally, the commodity image database constructed in this study is used to identify and verify the model. The recognition accuracy of the algorithm in this study is significantly higher than that of the Minitch stochastic gradient descent algorithm under different SNR conditions, and the recognition accuracy is the highest when SNR = 0 (99.3%). The test results show that the model proposed in this study has good recognition effect in the identification of commodity images in scenes of local occlusion, different perspectives, different backgrounds, and different light intensity, and the recognition accuracy is 97.1%. To sum up, the CIR platform based on MS-CNN model constructed in this study has high recognition accuracy and robustness, which can lay a foundation for the realization of subsequent intelligent commodity recognition technology.

Exploratory dynamics of vocal foraging during infant-caregiver communication

We investigated the hypothesis that infants search in an acoustic space for vocalisations that elicit adult utterances and vice versa, inspired by research on animal and human foraging. Infant-worn recorders were used to collect day-long audio recordings, and infant speech-related and adult vocalisation onsets and offsets were automatically identified. We examined vocalisation-to-vocalisation steps, focusing on inter-vocalisation time intervals and distances in an acoustic space defined by mean pitch and mean amplitude, measured from the child’s perspective. Infant inter-vocalisation intervals were shorter immediately following a vocal response from an adult. Adult intervals were shorter following an infant response and adult inter-vocalisation pitch differences were smaller following the receipt of a vocal response from the infant. These findings are consistent with the hypothesis that infants and caregivers are foraging vocally for social input. Increasing infant age was associated with changes in inter-vocalisation step sizes for both infants and adults, and we found associations between response likelihood and acoustic characteristics. Future work is needed to determine the impact of different labelling methods and of automatic labelling errors on the results. The study represents a novel application of foraging theory, demonstrating how infant behaviour and infant-caregiver interaction can be characterised as foraging processes.

Patient safety issues associated with the use of compounded medicines as alternatives to approved pharmaceutical products in Europe and how best practice can improve outcomes.

Pharmaceutical compounding allows individuals with special requirements access to medicines. Compounding can also be used to provide cheaper alternatives to commercially produced medicines which may be less strictly regulated than those commercially produced as they do not require marketing authorisation. This review describes the issues and potential risks associated with compounded medicines and equally importantly identifies best practices. To establish reports about lack of effectiveness, adverse events and medication errors occurring with compounded pharmaceuticals, a literature search was conducted of PubMed, Embase and MEDLINE databases for relevant cases in European countries which were published between 2003 and 2018. Case reports/series that described instances of successful use of compounded medicines over the same period were also identified. Overall, 12 case reports/case series describing problems associated with compounded medicines in Europe have been identified. Sources of patient risk associated with compounded medicines include lack of quality, safety and efficacy data, preparation and labelling errors, and improper storage and handling practices. Several case reports/series describing instances of overdose, medication errors and adverse events associated with compounded medications were reviewed. The number of affected patients was relatively small, but many were children and two adult patients experienced permanent sequelae. The number of incidents associated with compounded medicines is unknown, and so these numbers should be interpreted with caution. When licensed medicines are available, the use of compounded medicines can put patients at unnecessary risk which should be avoided. Stricter regulation is necessary to prevent similar cases from occurring in the future as the European market for compounded medications grows. Pharmacists can promote best practices in compounding through professional organisations. Future recommended actions are: 1. Stricter regulation is necessary to prevent similar cases from occurring in the future as the European market for compounded medications grows. 2. A comprehensive pan-European survey to gain a greater understanding of compounding procedures and techniques. This would provide valuable information to the benefit of hospital systems and their patients. 3. The results of the survey can then be used to improve the knowledge and quality control of compounded medicines for the good of patient safety.

Classifier‐Guided Visual Correction of Noisy Labels for Image Classification Tasks

AbstractTraining data plays an essential role in modern applications of machine learning. However, gathering labeled training data is time‐consuming. Therefore, labeling is often outsourced to less experienced users, or completely automated. This can introduce errors, which compromise valuable training data, and lead to suboptimal training results. We thus propose a novel approach that uses the power of pretrained classifiers to visually guide users to noisy labels, and let them interactively check error candidates, to iteratively improve the training data set. To systematically investigate training data, we propose a categorization of labeling errors into three different types, based on an analysis of potential pitfalls in label acquisition processes. For each of these types, we present approaches to detect, reason about, and resolve error candidates, as we propose measures and visual guidance techniques to support machine learning users. Our approach has been used to spot errors in well‐known machine learning benchmark data sets, and we tested its usability during a user evaluation. While initially developed for images, the techniques presented in this paper are independent of the classification algorithm, and can also be extended to many other types of training data.

Documentation errors in transfusion chain: Challenges and interventions

Background and objectiveThere are several steps in transfusion chain where accurate documentation is critical. This study was conducted to evaluate the frequency of documentation errors during transfusion process and to evaluate the effectiveness of interventions in error-management. Methods /MaterialThis study was conducted at Aga Khan University, Pakistan during 2016–2018. Transcription and bedside documentation errors were identified from in-house computerized system and from medical charts. Raw WBIT rate was calculated for repeat blood samples and adjusted for frequencies of ABO-groups in our population accounting for silent WBIT. Rate of ABO-mismatched red cell transfusions was calculated for the annual totals of red cell transfusions. Chi-square was used for observing relationship among errors of various data sets. ResultsA total of 43 WBIT was identified during 54,219 repeat blood samples where blood group was already defined in blood bank information system. Annual unadjusted and cryptic WBIT rate was consistent at 0.8 and 0.6 per 1000 samples respectively during 2016–2018 (p 0.859). There were 1161 transcription errors (1.1 %) in blood group documentation in 105,064 blood samples received for arranging blood products. ABO-mismatched transfusion rate was 0.9 for 10,000 RBC transfusions in pre- and decreased to 0.4 in post-typing era. Overall, the compliance for completing checklist, correct ABO technique and appropriate ABO-interpretation was 88 %, 40 % and 24 % in the reviewed medical charts. ConclusionsSample labeling errors were not improved through training or counseling. Bedside ABO-typing and checklist prior to blood transfusion can control the ABO-mismatched transfusion if done timely and correctly.

Rotulagem para alergênicos: uma avaliação dos rótulos comercializados com presença ou ausência de glúten e seus riscos inerentes a saúde dos celíacos no Brasil

A doença celíaca (DC) não é considerada uma alergia alimentar, trata-se de uma doença autoimune caracterizada por um processo inflamatório que envolve a mucosa do intestino delgado, em decorrência à ingestão de glúten, reduzindo assim a absorção dos nutrientes essenciais para o bom funcionamento do organismo. O único tratamento conhecido é a dieta isenta de glúten, adotada para toda a vida e deve ser rigorosa, completa, equilibrada e variada. Porém, no Brasil, ainda encontramos erros de rotulagem, circunstância grave, sendo que o rótulo é o modo que o celíaco encontra para se proteger do glúten e se há erro, será induzido a ingerir o antígeno tóxico, proporcionado o seu adoecimento e causando lesões corporais sérias. Sendo assim, o objetivo deste estudo é avaliar, através de uma revisão de literatura, o panorama da rotulagem de alimentos para celíacos no Brasil. Foram utilizadas como fonte de pesquisa as bases de dados da SCIELO, ANVISA, Ministério da Saúde, FENACELBRA, ASBAI, ACELBRA, MEDLINE, Google Acadêmico e livros, no período entre novembro e março de 2020. Logo, a pesquisa convalidou que a legislação deveria ser suficiente para promover um alinhamento na comunicação entre a indústria e o consumidor e que Vigilância Sanitária realiza intervenções para eliminar, diminuir ou prevenir riscos, porém observa-se que ainda é necessário a implementação de ações corretivas para que os portadores da DC tenham acesso a alimentação segura e que possam confiar nas palavras expressas nos rótulos das embalagens.

Monitoring and capturing patient identification errors in laboratory medicine

Specimen labelling and patient identification are significant contributors to the rate of error in the preanalytical phase of laboratory medicine. This study aimed to investigate the prevalence and nature of preanalytical quality monitoring practices for patient identification and specimen labelling errors in Irish clinical laboratories. A survey was developed by the Clinical Biochemistry Unit, Trinity College Dublin and the Irish External Quality Assessment Scheme (IEQAS), with the intention of gathering key information from each laboratory. Thirty-nine questions were organized into seven subsections covering general information, labelling requirements, information availability, rejection criteria, error monitoring, error reporting and interest in participation in an external quality assessment scheme. The survey was sent electronically to 63 laboratory quality managers at 55 laboratories in Ireland. A total of 39 responses (61% response rate) provided information on 94 separate laboratory departments or disciplines. Laboratories reported varying practices and requirements for labelling specimens and all accepted handwritten preprinted request forms. All (100%) respondents had defined rejection criteria both for specimen labelling and request form completion. Unsurprisingly, the rejection criteria differed between the various laboratory disciplines. Almost all respondents provided information to clinical staff on labelling requirements, but just over half provided training on the same. A large percentage of laboratories (74%) monitored the rate of specimen-labelling errors; however, only 46% had defined target limits for acceptable rates of error. The survey observed a wide variation in collection, recording and monitoring of errors but also confirmed significant interest in improving preanalytical monitoring and data collection.

Labeling compliance and species authentication of fish fillets sold at grocery stores in Southern California

Seafood mislabeling has numerous consequences, including economic deception and food safety risks. The focus of this study was to investigate fish species labeling, use of acceptable market names, and Country of Origin Labeling (COOL) compliance for fresh fish fillets sold at grocery store seafood counters in Southern California. A total of 120 fillets representing 16 different categories of fish were collected from 30 Perishable Agricultural Commodities Act (PACA)-listed grocery stores. Each sample underwent DNA barcoding to identify the species. Acceptable market names were confirmed using the FDA Seafood List. Samples were determined to be compliant with COOL if both the country of origin and the production method were declared in accordance with regulatory requirements. Species substitution was detected in 16 of the 120 samples (13.3%) and unacceptable market names were observed for an additional 11 samples (9.2%). The highest rates of species substitution were recorded for snapper (3/3), yellowtail (2/4), halibut (4/10), cod (3/10), and bass (2/7). COOL noncompliance was observed for 28 samples (23.3%): the country of origin was missing for 15 samples, production method was missing for 9 samples, and 4 samples were missing both. When all forms of mislabeling were considered, 47 of the 120 samples (39.2%) had at least one labeling error. The majority of grocery stores (25/30) had one or more samples with a mislabeling error. This study revealed species mislabeling as a continuous concern in the seafood industry, especially with higher-valued species. Furthermore, the lack of COOL compliance among retailers is concerning and suggests a need for increased focus on these regulations.

How Response Designs and Class Proportions Affect the Accuracy of Validation Data

Reference data collected to validate land-cover maps are generally considered free of errors. In practice, however, they contain errors despite best efforts to minimize them. These errors propagate during accuracy assessment and tweak the validation results. For photo-interpreted reference data, the two most widely studied sources of error are systematic incorrect labeling and vigilance drops. How estimation errors, i.e., errors intrinsic to the response design, affect the accuracy of reference data is far less understood. In this paper, we analyzed the impact of estimation errors for two types of classification systems (binary and multiclass) as well as for two common response designs (point-based and partition-based) with a range of sub-sample sizes. Our quantitative results indicate that labeling errors due to proportion estimations should not be neglected. They further confirm that the accuracy of response designs depends on the class proportions within the sampling units, with complex landscapes being more prone to errors. As a result, response designs where the number of sub-samples is predefined and fixed are inefficient. To guarantee high accuracy standards of validation data with minimum data collection effort, we propose a new method to adapt the number of sub-samples for each sample during the validation process. In practice, sub-samples are incrementally selected and labeled until the estimated class proportions reach the desired level of confidence. As a result, less effort is spent on labeling univocal cases and the spared effort can be allocated to more ambiguous cases. This increases the reliability of reference data and of subsequent accuracy assessment. Across our study site, we demonstrated that such an approach could reduce the labeling effort by 50% to 75%, with greater gains in homogeneous landscapes. We contend that adopting this optimization approach will not only increase the efficiency of reference data collection, but will also help deliver more reliable accuracy estimates to the user community.

Construction of Machine-Labeled Data for Improving Named Entity Recognition by Transfer Learning

Deep neural networks (DNNs) require a large amount of manually labeled training data to make significant achievements. However, manual labeling is laborious and costly. In this study, we propose a method for automatically generating training data and effectively using the generated data to reduce the labeling cost. The generated data (called “machine-labeled data”) is generated using a bagging-based bootstrapping approach. However, using the machine-labeled data does not guarantee high performance because of errors in the automatic labeling. In order to reduce the impact of mislabeling, we applied a transfer learning approach. The effect of our proposed method was verified with two versions of DNN-based named entity recognition (NER) models: bidirectional LSTM-CRF and vanilla BERT. We conducted NER tasks in two languages (English and Korean). The proposed method results in average F1 scores of 78.87% (3.9% point improvement) with bidirectional LSTM-CRF and 82.08% (1% point improvement) with BERT on three Korean NER datasets. In English, the performance increased by an average of 0.45% points with the two DNN-based models. The proposed NER systems outperform the baseline systems in both languages without the need for additional manual labeling.

Strengthening of banana breeding through data digitalization.

Improvement of edible bananas (a triploid and sterile crop) through conventional breeding is a challenging task owing to its recalcitrant nature for seed set, prolonged crop duration. In addition, the need of huge man power at different stages of progeny development and evaluation often leads to mislabeling, poor data management and loss of vital data. All this can be overcome by the application of advanced information technology source. This ensured secure and efficient data management such as storage, retrieval and data analysis and further could assist in tracking the breeding status in real time. Thus, a user-friendly web-based banana breeding tracker (BBT) has been developed using MySQL database with Hypertext Preprocessor (PHP). This BBT works on all operating systems with access to multiple users from anywhere at any time. Quick responsive (QR) code labels can be generated by the tracker, which can be decoded using QR scanner. Also for each and every updated progress in breeding stages, a new QR code can be generated, which in turn reduce labeling errors. Moreover, the tracker has additional tools to search, sort and filter the data from the data sets for efficient retrieval and analysis. This tracker is being upgraded with phenotypic and genotypic data that will be made available in the public domain for hastening the banana improvement program.

Automated cleaning of identity label noise in a large face dataset with quality control

For face recognition, some very large-scale datasets are publicly available in recent years, which are usually collected from the Internet using search engines, and thus have many faces with wrong identity (ID) labels (outliers). Additionally, the face images in these datasets have different qualities because of uncontrolled situations. The authors propose a novel approach for cleaning the ID label error, handling face images in different qualities. The face ID labels cleaned by their method can train better models for low-quality face recognition since more low-quality images are correctly labelled for training a deep model. In their low-to-high-quality face verification experiments, the deep model trained on their cleaning results of MS-Celeb-1M.v1 face dataset outperforms the same model trained on the same dataset cleaned by the semantic bootstrapping method. They also apply their ID label cleaning method on a subset of the cross-age celebrity dataset (CACD) face dataset, in which their quality-based cleaning can deliver higher precision and recall than a previous method on detecting the ID label errors.

A new strategy proposal to monitor ray fins landings in south‐east Brazil

Abstract Statistical fisheries data are usually obtained during landings, through rapid fish classification and their categorization under the trade names given to species or groups of species. However, species classification is often difficult, particularly concerning elasmobranchs whose fins and heads have been removed, leading to labelling errors. The aim of this paper is to identify the ray species composition landed in south‐east Brazil, providing identification strategies to support a plan for more efficient labelling and management. Samples were obtained from artisanal fleet landings, between November 2012 and May 2014. For taxonomic identification of whole or processed animals (n = 279, belonging to 10 species), morphological, metric, and molecular techniques were employed. The common name used by fishermen was more related to the location where they lived than the fishing gear used. Morphometric analyses resulted in pectoral fin to length and weight conversion equations of whole individuals for each species, and 10 variables were found to be diagnostic of each genus. A genetic identification, based on sequencing of the mitochondrial genes cytochrome b and cytochrome oxidase I, was applied to verify morphological identification. A dichotomous key that allowed ray identification at the species level from pectoral fin morphology was developed. The approach was field tested and deemed adequate, leading to a robust monitoring strategy for estimating the biomass of specific landed rays.

Building Experiential Teaching Content through the Perspective of Regression Standard of Living - Character language In Dept Reading Strategy of Character Language in Biography

Historical literature has both the nature of history and literature. The content is vast, the thoughts are rich. Plus, it has many values such as literature, history and philosophy. They also have accumulated a profound national culture. Various language teaching materials in the basic education stage are often carefully selected from the classic biography and the quality is fruitful. Learning these works while understanding history can enhance students' "construction and application of language", "development and promotion of thinking", "appreciation and creation of aesthetics" and "understanding and inheritance of culture” as core literacy has important meaning. However, there are misunderstandings in the teaching content of current historical works such as the teaching of historical essays, the narration of narrative analysis, the interpretation of historical reading, the lack of literary aesthetics, the analysis of facial analysis, the labelling and other errors directly lead to a boring and less interesting classroom teaching of historical works. Therefore, classroom teaching is difficult for students to produce affinity in Chinese history literature [1]. The classroom teaching of historical literature urgently needs to return to the regression standard of living in order to restore the historical life situation and construct experiential teaching content. Thus, only in this way can it be possible to build a foundation for understanding and let students walk into historical works. This article takes an excerpt from the "Records of the Grand Historian•Xiang Yu Exposition” in the "The Feast at Hong Gate” which stated, "he calls for me, a brotherhood was form" as an example, how to appreciate the character language by restoring the historical life situation with a life perspective.

Partial order relations for classification comparisons

AbstractThe Bayes classification rule offers the optimal classifier, minimizing the classification error rate, whereas the Neyman–Pearson lemma offers the optimal family of classifiers to maximize the detection rate for any given false alarm rate. These motivate studies on comparing classifiers based on similarities between the classifiers and the optimal. In this article, we define partial order relations on classifiers and families of classifiers, based on rankings of rate function values and rankings of test function values, respectively. Each partial order relation provides a sufficient condition, which yields better classification error rates or better performance on the receiver operating characteristic analysis. Various examples and applications of the partial order theorems are discussed to provide comparisons of classifiers and families of classifiers, including the comparison of cross‐validation methods, training data that contains outliers, and labelling errors in training data. The Canadian Journal of Statistics 48: 152–166; 2020 © 2019 Statistical Society of Canada

Corrigendum.

ImmunologyVolume 159, Issue 2 p. 242-242 CorrigendumFree Access Corrigendum This article corrects the following: CD6 monoclonal antibodies differ in epitope, kinetics and mechanism of action Lee I. Garner, Andrew Hartland, Johannes Breuning, Marion H. Brown, Volume 155Issue 2Immunology pages: 273-282 First Published online: June 13, 2018 First published: 14 November 2019 https://doi.org/10.1111/imm.13113AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat The article1 should be read with reference to new data2 for the correct specificity of the CD6 mAbs, OX124, OX125 and OX126.3, 4 Domain specificity for OX124, OX125 and OX126 should be deleted and the name of the antibody given. As the focus of the article is domain 1 mAbs, the overall conclusions are not altered. Specific alterations to the text are: Summary, p273. CD6 domain 1, OX125 and OX126 mAbs were equally effective in triggering interleukin-2 production…….CD6 domain 1 mAbs hindered binding of multivalent immobilised CD166 but were inferior compared with blocking by soluble CD166 or another CD6 mAb. Introduction, p274. Immunisation with soluble recombinant CD6 led to the production of novel CD6 mAbs(6). Materials and Methods, Monoclonal antibodies, p274. CD6, OX124 (mouse IgG1), OX125 (mouse IgG2b) and OX126 (mouse IgG1)(6) Materials and Methods, Flow cytometry, p275. CD6 mAb (OX126) Results, p277 and Figure 3. CD6 domain 1, OX125 and OX126 mAbs….. Results, p278. …CD6 mAbs, OX124, OX125 and OX126 specific for different epitopes on CD6 for efficacy …a CD6 mAb, OX125… interfere with OX126 binding to CD6. We began by asking do CD6 domain 1 mAbs hinder binding of OX126(6)……interactions of OX126. Results, p279 and Figure 4. CD6 mAb, OX126. Results, p279 and Figure 5. The data (Figure 5) are inconsistent with the specificity of OX1262 and indicate an antibody labelling error.3, 4 Read: CD6 domain 1 mAbs are less effective compared with another CD6 mAb….another CD6 mAb which was superior….another CD6 mAb was more effective >UMCD6>itolizumab at Discussion, p280. binding to domain 1 and by OX125 or OX126 in the efficacy of triggering.. Discussion, p281…competition experiment with a CD6 mAb, OX126..Delete text: “As none …soluble CD166(6)”. UMCD6 is not as efficacious at blocking CD6/CD166 interactions compared with another CD6 mAb References 1Garner LI, Hartland A, Breuning J, Brown MH. CD6 monoclonal antibodies differ in epitope, kinetics and mechanism of action. Immunology 2018; 155: 273– 82. 2Santos RF, Oliveira L, Brown MH, Carmo AM. Domain-specific CD6 monoclonal antibodies identify CD6 isoforms generated by alternative-splicing. Immunology 2019; 155: 273– 82. 3Hassan NJ, Simmonds SJ, Clarkson NG, Hanrahan S, Puklavec MJ, Bomb M et al. Second correction for Hassan et al., CD6 regulates T-cell responses through activation-dependent recruitment of the positive regulator SLP-76. Mol Cell Biol 2019; 39: e00054- 19. 4Hassan NJ, Simmonds SJ, Clarkson NG, Hanrahan S, Puklavec MJ, Bomb M et al. CD6 regulates T-cell responses through activation-dependent recruitment of the positive regulator SLP-76. Mol Cell Biol 2006; 26: 6727– 38. Volume159, Issue2February 2020Pages 242-242 ReferencesRelatedInformation