Association rule mining to identify transcription factor interactions in genomic regions

Gender Difference in Psychological, Cognitive, and Behavioral Patterns Among University Students During COVID-19: A Machine Learning Approach

The COVID-19 pandemic affects all population segments and is especially detrimental to university students because social interaction is critical for a rewarding campus life and valuable learning experiences. In particular, with the suspension of in-person activities and the adoption of virtual teaching modalities, university students face drastic changes in their physical activities, academic careers, and mental health. Our study applies a machine learning approach to explore the gender differences among U.S. university students in response to the global pandemic. Leveraging a proprietary survey dataset collected from 322 U.S. university students, we employ association rule mining (ARM) techniques to identify and compare psychological, cognitive, and behavioral patterns among male and female participants. To formulate our task under the conventional ARM framework, we model each unique question-answer pair of the survey questionnaire as a market basket item. Consequently, each participant's survey report is analogous to a customer's transaction on a collection of items. Our findings suggest that significant differences exist between the two gender groups in psychological distress and coping strategies. In addition, the two groups exhibit minor differences in cognitive patterns and consistent preventive behaviors. The identified gender differences could help professional institutions to facilitate customized advising or counseling for males and females in periods of unprecedented challenges.

Different Coping Patterns among US Graduate and Undergraduate Students during COVID-19 Pandemic: A Machine Learning Approach

COVID-19 caused unprecedented disruptions to regular university operations worldwide. Dealing with 100% virtual classrooms and suspension of essential in-person activities resulted in significant stress and anxiety for students coping with isolation, fear, and uncertainties in their academic careers. In this study, we applied a machine learning approach to identify distinct coping patterns between graduate and undergraduate students when facing these challenges. We based our study on a large proprietary dataset collected from 517 students in US professional institutions during an early peak of the pandemic. In particular, we cast our problem under the association rule mining (ARM) framework by introducing a new method to transform survey data into market basket items and customer transactions in which students' behavioral patterns were analogous to customer purchase patterns. Our experimental results suggested that graduate and undergraduate students adopted different ways of coping that could be attributed to their different maturity levels and lifestyles. Our findings can further serve as a focus of attention (FOA) tool to facilitate customized advising or counseling to address the unique challenges associated with each group that may warrant differentiated interventions.

Diagnosis and Treatment Rules of Chronic Kidney Disease and Nursing Intervention Models of Related Mental Diseases Using Electronic Medical Records and Data Mining

Objective

On the basis of electronic medical records, the data mining technology was adopted to explore the law of chronic kidney disease (CKD) and the intervention mode of mental health of patients.

Methods

Based on the electronic medical records, the corresponding data extraction, database establishment, and data cleaning of CKD were performed. After that, the related data analysis, frequency analysis, cluster analysis, and nonparametric analysis were used to explore the laws of CKD diagnosis and treatment and nursing intervention mode of mental illness. The most common causes of CKD were chronic glomerulonephritis (43.76%), aristolochic acid nephritis (16.34%), diabetic nephritis (12.87%), and hypertensive nephritis (11.58%). The major treatment method for end-stage patients was alternative therapies, accounting for 46%. Compared with the depression score before intervention, that of the patients after the mindfulness therapy (50.99 ± 9.77 vs. 47.01 ± 9.33, P=0.024 < 0.5) and target behaviour nursing intervention (52.21 ± 8.12 vs. 48.01 ± 9.33, P=0.032 < 0.05) was obviously decreased.

Conclusion

The data mining technology based on electronic records showed a good application prospect in the analysis of the diagnosis and treatment of CKD; and target behaviour nursing and mindfulness intervention were effective psychological intervention models.

Modelling the bioinformatics tertiary analysis research process

Background

With the advancements of Next Generation Techniques, a tremendous amount of genomic information has been made available to be analyzed by means of computational methods. Bioinformatics Tertiary Analysis is a complex multidisciplinary process that represents the final step of the whole bioinformatics analysis pipeline. Despite the popularity of the subject, the Bioinformatics Tertiary Analysis process has not yet been specified in a systematic way. The lack of a reference model results into a plethora of technological tools that are designed mostly on the data and not on the human process involved in Tertiary Analysis, making such systems difficult to use and to integrate.

Methods

To address this problem, we propose a conceptual model that captures the salient characteristics of the research methods and human tasks involved in Bioinformatics Tertiary Analysis. The model is grounded on a user study that involved bioinformatics specialists for the elicitation of a hierarchical task tree representing the Tertiary Analysis process. The outcome was refined and validated using the results of a vast survey of the literature reporting examples of Bioinformatics Tertiary Analysis activities.

Results

The final hierarchical task tree was then converted into an ontological representation using an ontology standard formalism. The results of our research provides a reference process model for Tertiary Analysis that can be used both to analyze and to compare existing tools, or to design new tools.

Conclusions

To highlight the potential of our approach and to exemplify its concrete applications, we describe a new bioinformatics tool and how the proposed process model informed its design.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12859-021-04310-5.

A self-attention model for inferring cooperativity between regulatory features

Deep learning has demonstrated its predictive power in modeling complex biological phenomena such as gene expression. The value of these models hinges not only on their accuracy, but also on the ability to extract biologically relevant information from the trained models. While there has been much recent work on developing feature attribution methods that discover the most important features for a given sequence, inferring cooperativity between regulatory elements, which is the hallmark of phenomena such as gene expression, remains an open problem. We present SATORI, a Self-ATtentiOn based model to detect Regulatory element Interactions. Our approach combines convolutional layers with a self-attention mechanism that helps us capture a global view of the landscape of interactions between regulatory elements in a sequence. A comprehensive evaluation demonstrates the ability of SATORI to identify numerous statistically significant TF-TF interactions, many of which have been previously reported. Our method is able to detect higher numbers of experimentally verified TF-TF interactions than existing methods, and has the advantage of not requiring a computationally expensive post-processing step. Finally, SATORI can be used for detection of any type of feature interaction in models that use a similar attention mechanism, and is not limited to the detection of TF-TF interactions.

Association Rule Mining and Prognostic Stratification of 2-Year Longevity in Octogenarians Undergoing Endovascular Therapy for Lower Extremity Arterial Disease: Observational Cohort Study

Background

Two-year longevity is a crucial consideration in revascularization strategies for patients with symptomatic lower extremity arterial disease (LEAD). However, factors associated with 2-year longevity and risk stratification in octogenarians or nonagenarians have been underreported.

Objective

This paper aims to investigate the associated variables and stratify the 2-year prognosis in older patients with LEAD.

Methods

We performed logistic regression and association rule mining based on the Apriori algorithm to discover independent variables and validate their associations with 2-year longevity. Malnutrition, inflammation, and stroke factors were identified. C statistics and Kaplan-Meier analysis were used to assess the impact of different numbers of malnutrition, inflammation, and stroke factors on 2-year longevity.

Results

We recruited a total of 232 octogenarians or nonagenarians (mean age 85 years, SD 4.2 years) treated with endovascular therapy. During the study period, 81 patients died, and 27 of those (33%) died from a cardiac origin within 2 years. Association rules analysis showed the interrelationships between 2-year longevity and the neutrophil-lymphocyte ratio (NLR) and nutritional status as determined by the Controlling Nutritional Status (CONUT) score or Geriatric Nutritional Risk Index (GNRI). The cut-off values of NLR, GNRI, and CONUT were ≥3.89, ≤90.3, and >3, respectively. The C statistics for the predictive power for 2-year longevity were similar between the CONUT score and the GNRI-based models (0.773 vs 0.760; P=.57). The Kaplan-Meier analysis showed that 2-year longevity was worse as the number of malnutrition, inflammation, and stroke factors increased from 0 to 3 in both the GNRI-based model (92% vs 68% vs 46% vs 12%, respectively; P<.001) and the CONUT score model (87% vs 75% vs 49% vs 10%, respectively; P<.001). The hazard ratio between those with 3 factors and those without was 18.2 (95% CI 7.0-47.2; P<.001) in the GNRI and 13.6 (95% CI 5.9-31.5; P<.001) in the CONUT score model.

Conclusions

This study demonstrated the association and crucial role of malnutrition, inflammation, and stroke factors in assessing 2-year longevity in older patients with LEAD. Using this simple risk score might assist clinicians in selecting the appropriate treatment.