Convolutional Neural Networks for Automatic Image Quality Control and EARL Compliance of PET Images

Background: Machine learning studies require a large number of images often obtained on different PET scanners. When merging these images, the use of harmonized images following EARL-standards is essential. However, when including retrospective images, EARL accreditation might not have been in place. The aim of this study was to develop a convolutional neural network (CNN) that can identify retrospectively if an image is EARL compliant and if it is meeting older or newer EARL-standards. Materials and Methods: 96 PET images acquired on three PET/CT systems were included in the study. All images were reconstructed with the locally clinically preferred, EARL1, and EARL2 compliant reconstruction protocols. After image pre-processing, one CNN was trained to separate clinical and EARL compliant reconstructions. A second CNN was optimized to identify EARL1 and EARL2 compliant images. The accuracy of both CNNs was assessed using 5-fold cross validation. The CNNs were validated on 24 images acquired on a PET scanner not included in the training data. To assess the impact of image noise on the CNN decision, the 24 images were reconstructed with different scan durations.Results: In the cross-validation, the first CNN classified all images correctly. When identifying EARL1 and EARL2 compliant images, the second CNN identified 100% EARL1 compliant and 85% EARL2 compliant images correctly. The accuracy in the independent dataset was comparable to the cross-validation accuracy. The scan duration had almost no impact on the results. Conclusion: The two CNNs trained in this study can be used to retrospectively include images in a multi-center setting by e.g. adding additional smoothing. This method is especially important for machine learning studies where the harmonization of images from different PET systems is essential.

LASP-1 interacts with ErbB2 in ovarian cancer cells

LASP-1 was identified as a protein following mass spectrometric analysis of phosphoproteins consequent to signaling by ErbB2 in SKOV-3 cells. It has been previously identified as an oncogene and is located on chromosomal arm 17q 0.76Mb centromeric to ErbB2. It is expressed in serous ovarian cancer cell lines as a 40kDa protein. In SKOV-3 cells, it was phosphorylated and was inhibited by Lapatinib and CP7274714. LASP-1 co-immunoprecipitated with ErbB2 in SKOV-3 cells, suggesting a direct interaction. This interaction and phosphorylation were independent of the kinase activity of ErbB2. Moreover, the binding of LASP-1 to ErbB2 was independent of the tyrosine phosphorylation of LASP-1. LASP-1 was neither expressed on the surface epithelium of the normal ovary nor in the fallopian tube. It was expressed in 28% of ovarian tumours (n=101) that did not significantly correlate with other clinical factors. In tumours from patients with invasive ductal carcinoma of the breast who had ErbB2 amplification (3+), LASP-1 was expressed in 3/20 (p <0.001). Analysis of the expression of an independent dataset of ovarian and breast tumors from TCGA showed the significant co-occurrence of ErbB2 and LASP-1 (p<0.01). These results suggest that LASP-1 and ErbB2 interaction could be important in the pathogenesis of ovarian cancer.

LMPred: Predicting Antimicrobial Peptides Using Pre-Trained Language Models and Deep Learning

Antimicrobial peptides (AMPs) are increasingly being used in the development of new therapeutic drugs, in areas such as cancer therapy and hypertension. Additionally, they are seen as an alternative to antibiotics due to the increasing occurrence of bacterial resistance. Wet-laboratory experimental identification, however, is both time consuming and costly, so in-silico models are now commonly used in order to screen new AMP candidates. This paper proposes a novel approach of creating model inputs; using pre-trained language models to produce contextualized embeddings representing the amino acids within each peptide sequence, before a convolutional neural network is then trained as the classifier. The optimal model was validated on two datasets, being one previously used in AMP prediction research, and an independent dataset, created by this paper. Predictive accuracies of 93.33% and 88.26% were achieved respectively, outperforming all previous state-of-the-art classification models.

Musical Mix Clarity Prediction Using Decomposition and Perceptual Masking Thresholds

Objective measurement of perceptually motivated music attributes has application in both target-driven mixing and mastering methodologies and music information retrieval. This work proposes a perceptual model of mix clarity which decomposes a mixed input signal into transient, steady-state, and residual components. Masking thresholds are calculated for each component and their relative relationship is used to determine an overall masking score as the model’s output. Three variants of the model were tested against subjective mix clarity scores gathered from a controlled listening test. The best performing variant achieved a Spearman’s rank correlation of rho = 0.8382 (p < 0.01). Furthermore, the model output was analysed using an independent dataset generated by progressively applying degradation effects to the test stimuli. Analysis of the model suggested a close relationship between the proposed model and the subjective mix clarity scores particularly when masking was measured using linearly spaced analysis bands. Moreover, the presence of noise-like residual signals was shown to have a negative effect on the perceived mix clarity.

Convolutional Neural Networks for automatic image quality control and EARL compliance of PET images

Background Machine learning studies require a large number of images often obtained on different PET scanners. When merging these images, the use of harmonized images following EARL-standards is essential. However, when including retrospective images, EARL accreditation might not have been in place. The aim of this study was to develop a convolutional neural network (CNN) that can identify retrospectively if an image is EARL compliant and if it is meeting older or newer EARL-standards. Materials and Methods 96 PET images acquired on three PET/CT systems were included in the study. All images were reconstructed with the locally clinically preferred, EARL1, and EARL2 compliant reconstruction protocols. After image pre-processing, one CNN was trained to separate clinical and EARL compliant reconstructions. A second CNN was optimized to identify EARL1 and EARL2 compliant images. The accuracy of both CNNs was assessed using 5-fold cross validation. The CNNs were validated on 24 images acquired on a PET scanner not included in the training data. To assess the impact of image noise on the CNN decision, the 24 images were reconstructed with different scan durations. Results In the cross-validation, the first CNN classified all images correctly. When identifying EARL1 and EARL2 compliant images, the second CNN identified 100% EARL1 compliant and 85% EARL2 compliant images correctly. The accuracy in the independent dataset was comparable to the cross-validation accuracy. The scan duration had almost no impact on the results. Conclusion The two CNNs trained in this study can be used to retrospectively include images in a multi-center setting by e.g. adding additional smoothing. This method is especially important for machine learning studies where the harmonization of images from different PET systems is essential.

BOLD Coupling between Lesioned and Healthy Brain Is Associated with Glioma Patients’ Recovery

Predicting functional outcomes after surgery and early adjuvant treatment is difficult due to the complex, extended, interlocking brain networks that underpin cognition. The aim of this study was to test glioma functional interactions with the rest of the brain, thereby identifying the risk factors of cognitive recovery or deterioration. Seventeen patients with diffuse non-enhancing glioma (aged 22–56 years) were longitudinally MRI scanned and cognitively assessed before and after surgery and during a 12-month recovery period (55 MRI scans in total after exclusions). We initially found, and then replicated in an independent dataset, that the spatial correlation pattern between regional and global BOLD signals (also known as global signal topography) was associated with tumour occurrence. We then estimated the coupling between the BOLD signal from within the tumour and the signal extracted from different brain tissues. We observed that the normative global signal topography is reorganised in glioma patients during the recovery period. Moreover, we found that the BOLD signal within the tumour and lesioned brain was coupled with the global signal and that this coupling was associated with cognitive recovery. Nevertheless, patients did not show any apparent disruption of functional connectivity within canonical functional networks. Understanding how tumour infiltration and coupling are related to patients’ recovery represents a major step forward in prognostic development.

Machine learning based decipherment of Cell Population Data: a promising hospital front-door screening tool for COVID-19

Introduction/Objective Key challenges against early diagnosis of COVID-19 are its symptoms sharing nature and prolong SARS-CoV-2 PCR turnaround time. Hither machine learning (ML) tools experienced by routinely generated clinical data; potentially grant early prediction. Methods/Case Report Routine and earlier diagnostic data along demographic information were extracted for total of 21,672 subsequent presentations. Along conventional statistics, multilayer perceptron (MLP) and radial basis function (RBF) were applied to predict COVID-19 from pre-pandemic control. Three feature sets were prepared, and performance evaluated through stratified 10-fold cross validation. With differing predominance of COVID-19, multiple test sets were created and predictive efficiency was evaluated to simulate real-fashion performance against fluctuating course of pandemic. Models validation was also inducted in prospective manner on independent dataset, equating framework forecasting to conclusions from PCR. Results (if a Case Study enter NA) RBF model attained superior cross entropy error 20.761(7.883) and 20.782(3.991) for Q-Flags and Routine Items respectively while MLP outperformed for cell population data (CPD) parameters with value of 6.968(1.259) for ‘training(testing)’. Our CPD driven MLP framework in challenge of lower (&lt;5%) COVID-19 predominance affords greater negative predictive values (NPV &gt;99%). Higher accuracy (%correct 92.5) was offered during prospective validation using independent dataset. Sensitivity analysis advances illusive accuracy (%correct 94.1) and NPV (96.9%). LY-WZ, Blasts/Abn Lympho?, ‘HGB Interf?’, and ‘RBC Agglutination?’ are among novel enlightening study attributes. Conclusion CPD driven ML tools offer efficient screening of COVID-19 patients at presentation to hospital to backing early expulsion and directing patients’ flow-from amid the initial presentation to hospital.

IL13Pred: A method for predicting immunoregulatory cytokine IL-13 inducing peptides for managing COVID-19 severity

Interleukin 13 (IL-13) is an immunoregulatory cytokine that is primarily released by activated T-helper 2 cells. It induces the pathogenesis of many allergic diseases, such as airway hyperresponsiveness, glycoprotein hypersecretion and goblet cell hyperplasia. IL-13 also inhibits tumor immunosurveillance, which leads to carcinogenesis. In recent studies, elevated IL-13 serum levels have been shown in severe COVID-19 patients. Thus it is important to predict IL-13 inducing peptides or regions in a protein for designing safe protein therapeutics particularly immunotherapeutic. This paper describes a method developed for predicting, designing and scanning IL-13 inducing peptides. The dataset used in this study contain experimentally validated 313 IL-13 inducing peptides and 2908 non-inducing homo-sapiens peptides extracted from the immune epitope database (IEDB). We have extracted 95 key features using SVC-L1 technique from the originally generated 9165 features using Pfeature. Further, these key features were ranked based on their prediction ability, and top 10 features were used for building machine learning prediction models. In this study, we have deployed various machine learning techniques to develop models for predicting IL-13 inducing peptides. These models were trained, test and evaluated using five-fold cross-validation techniques; best model were evaluated on independent dataset. Our best model based on XGBoost achieves a maximum AUC of 0.83 and 0.80 on the training and independent dataset, respectively. Our analysis indicate that certain SARS-COV2 variants are more prone to induce IL-13 in COVID-19 patients. A standalone package as well as a web server named ‘IL-13Pred’ has been developed for predicting IL-13 inducing peptides (https://webs.iiitd.edu.in/raghava/il13pred/).Key PointsInterleukin-13, an immunoregulatory cytokine plays an important role in increasing severity of COVID-19 and other diseases.IL-13Pred is a highly accurate in-silico method developed for predicting the IL-13 inducing peptides/ epitopes.IL-13 inducing peptides are reported in various SARS-CoV2 strains/variants proteins.This method can be used to detect IL-13 inducing peptides in vaccine candidates.User friendly web server and standalone software is freely available for IL-13PredAuthor’s BiographyShipra Jain is currently working as Ph.D. in Computational Biology from Department of Computational Biology, Indraprastha Institute of Information Technology, New Delhi, India.Anjali Dhall is currently working as Ph.D. in Computational Biology from Department of Computational Biology, Indraprastha Institute of Information Technology, New Delhi, India.Sumeet Patiyal is currently working as Ph.D. in Computational Biology from Department of Computational Biology, Indraprastha Institute of Information Technology, New Delhi, India.Gajendra P. S. Raghava is currently working as Professor and Head of Department of Computational Biology, Indraprastha Institute of Information Technology, New Delhi, India.

Generalizing effects of frontostriatal structural connectivity on self-esteem using predictive modeling

Determining the generalizability of biological mechanisms sup- porting psychological constructs is a central goal of cognitive neuroscience. Self-esteem is a popular psychological construct that is associated with a variety of measures of mental health and life satisfaction. Recently, there has been interest in identifying biological mechanisms that support individual differences in self-esteem. Understanding the biological basis of self-esteem requires identifying predictive biomarkers of self-esteem that generalize across groups of individuals. Previous research us- ing diffusion magnetic resonance imaging has shown that self- esteem is related to the integrity of structural connections link- ing frontostriatal brain systems involved in self-referential processing and reward. However, these findings were based on a small, relatively homogeneous group of participants. In the cur- rent study, we used an out-of-sample predictive modeling approach to generalize the results of the previous study to an independent sample of participants more than twice the size of the original study. We found that both linear univariate and multivariate machine learning models trained on frontostriatal integrity from the original data significantly predicted self-esteem in the independent dataset. These findings underscore the relationship between self-esteem and frontostriatal connectivity and suggest these results are robust to differences in scanning acquisition, analytic methods, and participant demographics.