Predicting peptide properties from mass spectrometry data using deep attention-based multitask network and uncertainty quantification.

Database search algorithms reduce the number of potential candidate peptides against which scoring needs to be performed using a single (i.e. mass) property for filtering. While useful, filtering based on one property may lead to exclusion of non-abundant spectra and uncharacterized peptides - potentially exacerbating the streetlight effect. Here we present ProteoRift, a novel attention and multitask deep-network, which can predict multiple peptide properties (length, missed cleavages, and modification status) directly from spectra. We demonstrate that ProteoRift can predict these properties with up to 97% accuracy resulting in search-space reduction by more than 90%. As a result, our end-to-end pipeline is shown to exhibit 8x to 12x speedups with peptide deduction accuracy comparable to algorithmic techniques. We also formulate two uncertainty estimation metrics, which can distinguish between in-distribution and out-of-distribution data (ROC-AUC 0.99) and predict high-scoring mass spectra against correct peptide (ROC-AUC 0.94). These models and metrics are integrated in an end-to-end ML pipeline available at https://github.com/pcdslab/ProteoRift.

Palliative Care Acceptability for Older Adults with Advanced CKD: A Qualitative Study of Patients and Nephrologists.

Rationale & Objective: Older adults in the United States often receive kidney therapies that do not align with their goals. Palliative care (PC) specialists are experts in assisting patients with the goals of care discussions and decision support, yet views and experiences of older patients who have received PC while contemplating kidney therapy decisions and their nephrologists remain unexplored. We evaluated the acceptability of CKD-EDU, a PC-based kidney therapy decision support intervention for adults ≥75 years of age. Study Design: Qualitative study. Setting & Participants: Two trained research coordinators interviewed patients and nephrologists participating in the CKD-EDU study. Analytical Approach: Three coders analyzed the qualitative data using a thematic analysis approach to identify salient themes pertaining to intervention acceptability. Results: Patients (n = 19; mean age: 80 years) viewed the PC intervention favorably, noting PC physicians' excellent communication skills, whole-person care, and decision-making support, including comprehension of prognostic information. Nephrologists (n = 24; mean age) welcomed PC assistance in decision making, support for conservative kidney management, and symptom management; a minority voiced concerns about third-party involvement in their practice. Limitations: Single-center study. Conclusions: Overall, patients and nephrologists generally found the PC intervention to be acceptable. Future testing of the current PC-based decision support intervention in a larger randomized controlled trial for older people navigating kidney therapy decisions is needed.

Literature on the acceptability of palliative care for kidney therapy decision making for older adults is scarce. This qualitative study establishes the acceptability of a palliative care (PC)-based kidney therapy decision support pilot intervention among older adults with advanced chronic kidney disease (CKD). Both patients and nephrologists found the intervention acceptable. Future testing of this PC-based intervention in an adequately powered randomized controlled trial for older individuals navigating kidney therapy decisions is needed.

Comparing Nephrologists' Self-Reported Decision-Making Skills and Treatment Attitudes With Their Patients' Experiences of Making Kidney Therapy Decisions and Receiving Nephrology Care.

BACKGROUND: Dialysis is often initiated in the United States without exploring patients' preferred decision-making style, and conservative kidney management (CKM) is infrequently presented. To improve kidney therapy (KT) decision-making, research on nephrologists' comfort with various decision-making styles, attitudes towards CKM, and reports of patients' lived experiences with KT decision-making is needed. METHODS: We surveyed 28 nephrologists and 58 of their patients aged ≥75 years. The nephrologist survey was designed to gauge their comfort levels with decision-making styles and attitudes towards CKM. The patient survey assessed experiences in making KT decisions. RESULTS: The average age of nephrologists was 43 years, and that of patients was 82 years. Nephrologists rated themselves as comfortable with various decision styles: paternalistic (60.7%), shared decision-making (92.8%), and patient-driven decision-making (67.8%). Nearly 57% of nephrologists felt challenged or were neutral in determining CKM's suitability, and 39% reported difficulties in discussing CKM with patients or were neutral. Only 38 % of patients recalled discussing CKM with their nephrologists, and a minority reported discussing CKM-related topics such as life expectancy (24.7%), quality of life (QOL) (45.1%), and end-of-life care (17.5%). CONCLUSIONS: Most nephrologists displayed comfort with various decision-making styles; however, many described difficulties in guiding patients toward CKM. In contrast, patients reported gaps in vital aspects of KT decision-making and CKM choices, such as discussions of life expectancy, QOL, and end-of-life care. Raising awareness of blind spots in decision-making skills and educating nephrologists in KT decision-making to include CKM and other person-centered aspects of care are needed.

How can event attribution science underpin financial decisions on Loss and Damage?

With climate extremes hitting nations across the globe, disproportionately burdening vulnerable developing countries, the prompt operation of the Loss and Damage fund is of paramount importance. As decisions on resource disbursement at the international level, and investment strategies at the national level, loom, the climate science community's role in providing fair and effective evidence is crucial. Attribution science can provide useful information for decision makers, but both ethical implications and deep uncertainty cannot be ignored. Considering these aspects, we articulate a vision that integrates established attribution methods and multiple lines of evidence within a coherent logical framework.

Making MS Omics Data ML-Ready: SpeCollate Protocols.

The increasing complexity and volume of mass spectrometry (MS) data have presented new challenges and opportunities for proteomics data analysis and interpretation. In this chapter, we provide a comprehensive guide to transforming MS data for machine learning (ML) training, inference, and applications. The chapter is organized into three parts. The first part describes the data analysis needed for MS-based experiments and a general introduction to our deep learning model SpeCollate-which we will use throughout the chapter for illustration. The second part of the chapter explores the transformation of MS data for inference, providing a step-by-step guide for users to deduce peptides from their MS data. This section aims to bridge the gap between data acquisition and practical applications by detailing the necessary steps for data preparation and interpretation. In the final part, we present a demonstrative example of SpeCollate, a deep learning-based peptide database search engine that overcomes the problems of simplistic simulation of theoretical spectra and heuristic scoring functions for peptide-spectrum matches by generating joint embeddings for spectra and peptides. SpeCollate is a user-friendly tool with an intuitive command-line interface to perform the search, showcasing the effectiveness of the techniques and methodologies discussed in the earlier sections and highlighting the potential of machine learning in the context of mass spectrometry data analysis. By offering a comprehensive overview of data transformation, inference, and ML model applications for mass spectrometry, this chapter aims to empower researchers and practitioners in leveraging the power of machine learning to unlock novel insights and drive innovation in the field of mass spectrometry-based omics.

Uncovering Patient and Caregiver Goals for Goal-Concordant Care in Kidney Therapy Decisions.

CONTEXT: In kidney therapy (KT) decisions, goal-concordant decision-making is recognized to be important, yet alignment with patients' goals during dialysis initiation is not always achieved. OBJECTIVES: To explore older patients' and caregivers' hopes, goals, and fears related to KT and communication of these elements with members of their health care team. METHODS: The study included patients aged ≥75 years with an estimated glomerular filtration rate ≤25 mL/min/1.73 m2 and their caregivers enrolled in a palliative care intervention for KT decision-making. Patients and caregivers were asked open-ended questions about their hopes, goals, and fears related to KT decisions. A survey assessed if patients shared their goals with members of their health care team. Qualitative data underwent content analysis, supplemented by demographic descriptive statistics. RESULTS: The mean age of patients (n = 26) was 82.7 (±5.7) years, and caregivers (n = 15) had a mean age of 66.4 (±13.7) years. Among the participants, 13 patients and 11 caregivers were women, and 20 patients and 12 caregivers were White. Four themes emerged: (1) Maintaining things as good as they are by avoiding dialysis-related burdens; (2) seeking longevity while avoiding dialysis; (3) avoiding pain, symptoms, and body disfigurement; and (4) deferring decision-making. Patients rarely had shared their goals with the key members of their health care team. CONCLUSION: Patients and caregivers prioritize maintaining quality of life, deferring decision-making regarding dialysis, and avoiding dialysis-related burdens. These goals are often unshared with their family and health care teams. Given our aging population, urgent action is needed to educate clinicians to actively explore and engage with patient goals in KT decision-making.

PVTAD: ALZHEIMER'S DISEASE DIAGNOSIS USING PYRAMID VISION TRANSFORMER APPLIED TO WHITE MATTER OF T1-WEIGHTED STRUCTURAL MRI DATA.

Alzheimer's disease (AD) is a neurodegenerative disorder, and timely diagnosis is crucial for early interventions. AD is known to have disruptive local and global brain neural connections that may be instrumental in understanding and extracting specific biomarkers. Previous machine-learning approaches are mostly based on convolutional neural network (CNN) and standard vision transformer (ViT) models which may not sufficiently capture the multidimensional local and global patterns that may be indicative of AD. Therefore, in this paper, we propose a novel approach called PVTAD to classify AD and cognitively normal (CN) cases using pretrained pyramid vision transformer (PVT) and white matter (WM) of T1-weighted structural MRI (sMRI) data. Our approach combines the advantages of CNN and standard ViT to extract both local and global features indicative of AD from the WM coronal middle slices. We performed experiments on subjects with T1-weighed MPRAGE sMRI scans from the ADNI dataset. Our results demonstrate that the PVTAD achieves an average accuracy of 97.7% and F1-score of 97.6%, outperforming the single and parallel CNN and standard ViT architectures based on sMRI data for AD vs. CN classification.

GPU-acceleration of the distributed-memory database peptide search of mass spectrometry data.

Database peptide search is the primary computational technique for identifying peptides from the mass spectrometry (MS) data. Graphical Processing Units (GPU) computing is now ubiquitous in the current-generation of high-performance computing (HPC) systems, yet its application in the database peptide search domain remains limited. Part of the reason is the use of sub-optimal algorithms in the existing GPU-accelerated methods resulting in significantly inefficient hardware utilization. In this paper, we design and implement a new-age CPU-GPU HPC framework, called GiCOPS, for efficient and complete GPU-acceleration of the modern database peptide search algorithms on supercomputers. Our experimentation shows that the GiCOPS exhibits between 1.2 to 5[Formula: see text] speed improvement over its CPU-only predecessor, HiCOPS, and over 10[Formula: see text] improvement over several existing GPU-based database search algorithms for sufficiently large experiment sizes. We further assess and optimize the performance of our framework using the Roofline Model and report near-optimal results for several metrics including computations per second, occupancy rate, memory workload, branch efficiency and shared memory performance. Finally, the CPU-GPU methods and optimizations proposed in our work for complex integer- and memory-bounded algorithmic pipelines can also be extended to accelerate the existing and future peptide identification algorithms. GiCOPS is now integrated with our umbrella HPC framework HiCOPS and is available at: https://github.com/pcdslab/gicops .

Confounding Effects on the Performance of Machine Learning Analysis of Static Functional Connectivity Computed from rs-fMRI Multi-site Data.

Resting-state functional magnetic resonance imaging (rs-fMRI) is a non-invasive imaging technique widely used in neuroscience to understand the functional connectivity of the human brain. While rs-fMRI multi-site data can help to understand the inner working of the brain, the data acquisition and processing of this data has many challenges. One of the challenges is the variability of the data associated with different acquisitions sites, and different MRI machines vendors. Other factors such as population heterogeneity among different sites, with variables such as age and gender of the subjects, must also be considered. Given that most of the machine-learning models are developed using these rs-fMRI multi-site data sets, the intrinsic confounding effects can adversely affect the generalizability and reliability of these computational methods, as well as the imposition of upper limits on the classification scores. This work aims to identify the phenotypic and imaging variables producing the confounding effects, as well as to control these effects. Our goal is to maximize the classification scores obtained from the machine learning analysis of the Autism Brain Imaging Data Exchange (ABIDE) rs-fMRI multi-site data. To achieve this goal, we propose novel methods of stratification to produce homogeneous sub-samples of the 17 ABIDE sites, as well as the generation of new features from the static functional connectivity values, using multiple linear regression models, ComBat harmonization models, and normalization methods. The main results obtained with our statistical models and methods are an accuracy of 76.4%, sensitivity of 82.9%, and specificity of 77.0%, which are 8.8%, 20.5%, and 7.5% above the baseline classification scores obtained from the machine learning analysis of the static functional connectivity computed from the ABIDE rs-fMRI multi-site data.

Development and Acceptability of a Kidney Therapy Decision Aid for Patients Aged 75 Years and Older: A Design-Based Research Involving Patients, Caregivers, and a Multidisciplinary Team.

Rationale & Objective: Many older adults prefer quality of life over longevity, and some prefer conservative kidney management (CKM) over dialysis. There is a lack of patient-decision aids for adults aged 75 years or older facing kidney therapy decisions, which not only include information on dialysis and CKM but also encourage end-of-life planning. We iteratively developed a paper-based patient-decision aid for older people with low literacy and conducted surveys to assess its acceptability. Study Design: Design-based research. Setting and Participants: Informed by design-based research principles and theory of behavioral activation, a multidisciplinary team of experts created a first version of the patient-decision aid containing 2 components: (1) educational material about kidney therapy options such as CKM, and (2) a question prompt list relevant to kidney therapy and end-of-life decision making. On the basis of the acceptability input of patients and caregivers, separate qualitative interviews of 35 people receiving maintenance dialysis, and with the independent feedback of educated layperson, we further modified the patient-decision aid to create a second version. Analytical Approach: We used descriptive statistics to present the results of acceptability surveys and thematic content analyses for patients' qualitative interviews. Results: The mean age of patients (n=21) who tested the patient-decision aid was 80 years and the mean age of caregivers (n=9) was 70 years. All respondents held positive views about the educational component and would recommend the educational component to others (100% patients and caregivers). Most of the patients reported that the question prompt list helped them put concerns into words (80% patients and 88% caregivers) and would recommend the question prompt list to others (95% patients and 100% caregivers). Limitations: Single-center study. Conclusions: Both components of the patient-decision aid received high acceptability ratings. We plan to launch a larger effectiveness study to test the outcomes of a decision-supporting intervention combining the patient-decision aid with palliative care-based decision coaching.

PPAD: a deep learning architecture to predict progression of Alzheimer's disease.

MOTIVATION: Alzheimer's disease (AD) is a neurodegenerative disease that affects millions of people worldwide. Mild cognitive impairment (MCI) is an intermediary stage between cognitively normal state and AD. Not all people who have MCI convert to AD. The diagnosis of AD is made after significant symptoms of dementia such as short-term memory loss are already present. Since AD is currently an irreversible disease, diagnosis at the onset of the disease brings a huge burden on patients, their caregivers, and the healthcare sector. Thus, there is a crucial need to develop methods for the early prediction AD for patients who have MCI. Recurrent neural networks (RNN) have been successfully used to handle electronic health records (EHR) for predicting conversion from MCI to AD. However, RNN ignores irregular time intervals between successive events which occurs common in electronic health record data. In this study, we propose two deep learning architectures based on RNN, namely Predicting Progression of Alzheimer's Disease (PPAD) and PPAD-Autoencoder. PPAD and PPAD-Autoencoder are designed for early predicting conversion from MCI to AD at the next visit and multiple visits ahead for patients, respectively. To minimize the effect of the irregular time intervals between visits, we propose using age in each visit as an indicator of time change between successive visits. RESULTS: Our experimental results conducted on Alzheimer's Disease Neuroimaging Initiative and National Alzheimer's Coordinating Center datasets showed that our proposed models outperformed all baseline models for most prediction scenarios in terms of F2 and sensitivity. We also observed that the age feature was one of top features and was able to address irregular time interval problem. AVAILABILITY AND IMPLEMENTATION: https://github.com/bozdaglab/PPAD.

Description of Dissolved Organic Matter Transformational Networks at the Molecular Level.

Dissolved Organic Matter (DOM) is an important component of the global carbon cycle. Unscrambling the structural footprint of DOM is key to understand its biogeochemical transformations at the mechanistic level. Although numerous studies have improved our knowledge of DOM chemical makeup, its three-dimensional picture remains largely unrevealed. In this work, we compare four solid phase extracted (SPE) DOM samples from three different freshwater ecosystems using high resolution mobility and ultrahigh-resolution Fourier transform ion cyclotron resonance tandem mass spectrometry (FT-ICR MS/MS). Structural families were identified based on neutral losses at the level of nominal mass using continuous accumulation of selected ions-collision induced dissociation (CASI-CID)FT-ICR MS/MS. Comparison of the structural families indicated dissimilarities in the structural footprint of this sample set. The structural family representation using Cytoscape software revealed characteristic clustering patterns among the DOM samples, thus confirming clear differences at the structural level (Only 10% is common across the four samples.). The analysis at the level of neutral loss-based functionalities suggests that hydration and carboxylation are ubiquitous transformational processes across the three ecosystems. In contrast, transformation mechanisms involving methoxy moieties may be constrained in estuarine systems due to extensive upstream lignin biodegradation. The inclusion of the isomeric content (mobility measurements at the level of chemical formula) in the structural family description suggests that additional transformation pathways and/or source variations are possible and account for the dissimilarities observed. While the structural character of more and diverse types of DOM samples needs to be assessed and added to this database, the results presented here demonstrate that Graph-DOM is a powerful tool capable of providing novel information on the DOM chemical footprint, based on structural interconnections of precursor molecules generated by fragmentation pathways and collisional cross sections.

PPAD: A deep learning architecture to predict progression of Alzheimer's disease.

Alzheimer's disease (AD) is a neurodegenerative disease that affects millions of people worldwide. Mild cognitive impairment (MCI) is an intermediary stage between cognitively normal (CN) state and AD. Not all people who have MCI convert to AD. The diagnosis of AD is made after significant symptoms of dementia such as short-term memory loss are already present. Since AD is currently an irreversible disease, diagnosis at the onset of disease brings a huge burden on patients, their caregivers, and the healthcare sector. Thus, there is a crucial need to develop methods for the early prediction AD for patients who have MCI. Recurrent Neural Networks (RNN) have been successfully used to handle Electronic Health Records (EHR) for predicting conversion from MCI to AD. However, RNN ignores irregular time intervals between successive events which occurs common in EHR data. In this study, we propose two deep learning architectures based on RNN, namely Predicting Progression of Alzheimer's Disease (PPAD) and PPAD-Autoencoder (PPAD-AE). PPAD and PPAD-AE are designed for early predicting conversion from MCI to AD at the next visit and multiple visits ahead for patients, respectively. To minimize the effect of the irregular time intervals between visits, we propose using age in each visit as an indicator of time change between successive visits. Our experimental results conducted on Alzheimer's Disease Neuroimaging Initiative (ADNI) and National Alzheimer's Coordinating Center (NACC) datasets showed that our proposed models outperformed all baseline models for most prediction scenarios in terms of F2 and sensitivity. We also observed that the age feature was one of top features and was able to address irregular time interval problem.

Racial Disparities in Health Beliefs and Advance Care Planning Among Patients Receiving Maintenance Dialysis.

CONTEXT: Among people receiving maintenance dialysis, little is known about racial disparities in the occurrence of prognostic discussions, beliefs about future health, and completion of advance care planning (ACP) documents. OBJECTIVES: We examined whether Black patients receiving maintenance dialysis differ from White patients in prognostic discussions, beliefs about future health, and completion of ACP-related documents. METHODS: We surveyed adult patients receiving maintenance dialysis from seven dialysis units in Cleveland, Ohio, and hospitalized patients at a tertiary care hospital in Cleveland. Of the 450 patients who were asked to participate in the study, 423 (94%) agreed. We restricted the current secondary analyses to include only Black (n=285) and White (n=114) patients. The survey assessed patients' knowledge of their kidney disease, attitudes toward chronic kidney disease (CKD) treatment, preferences for end-of-life (EoL) care, the patient-reported occurrence of prognostic discussions, experiences with kidney therapy decision making, sentiments of dialysis regret, beliefs about health over the next 12 months, and advance care planning. We used stepwise logistic regression to determine if race was associated with the occurrence of prognostic discussions, beliefs about future health, and completion of an ACP-related document, while controlling for potential confounders. RESULTS: We found no significant difference in the frequency of prognostic discussions between Black (11.9%) versus White patients (7%) (P=0.15). However, Black patients (19%) had lower odds of believing that their health would worsen over the next 12 months (OR 0.22, CI 0.12, 0.44) and reporting completion of any ACP-related document (OR 0.5, CI 0.32, 0.81) compared to White patients CONCLUSION: Racial differences exist in beliefs about future health and completion of ACP-related documents. Systemic efforts to investigate differences in health beliefs and address racial disparities in the completion of ACP-related documents are needed.

ASD-GResTM: Deep Learning Framework for ASD classification using Gramian Angular Field.

Autism Spectrum Disorder (ASD) is a heterogeneous disorder in children, and the current clinical diagnosis is accomplished using behavioral, cognitive, developmental, and language metrics. These clinical metrics can be imperfect measures as they are subject to high test-retest variability, and are influenced by assessment factors such as environment, social structure, or comorbid disorders. Advances in neuroimaging coupled with machine-learning provides an opportunity to develop methods that are more quantifiable, and reliable than existing clinical techniques. In this paper, we design and develop a deep-learning model that operates on functional magnetic resonance imaging (fMRI) data, and can classify between ASD and neurotypical brains. We introduce a novel strategy to transform time-series data extracted from fMRI signals into Gramian Angular Field (GAF) while locking in the temporal and spatial patterns in the data. Our motivation is to design and develop a novel framework that could encode the time-series, acquired from fMRI data, into images that can be used by deep-learning architectures that have been successful in computer vision. In our proposed framework called ASD-GResTM, we used a Convolutional Neural Network (CNN) to extract useful features from GAF images. We then used a Long Short-Term Memory (LSTM) layer to learn the activities between the regions. Finally, the output representations of the last LSTM layer are applied to a single-layer perceptron (SPL) to get the final classification. Our extensive experimentation demonstrates high accuracy across 4 centers, and outperforms state-of-the-art models on two centers with an increase in the accuracy of 17.58% and 6.7%, respectively as compared to the state of the art. Our model achieved the maximum accuracy of 81.78% with high degree of sensitivity and specificity. All training, validation, and testing was accomplished using openly available ABIDE-I benchmarking dataset.