A longitudinal feature selection method identifies relevant genes to distinguish complicated injury and uncomplicated injury over time

Muscle Atrophy Evaluation via Radiomics Analysis using Ultrasound Images: A Cohort Data Study

OBJECTIVE

Existing methods for muscle atrophy evaluation based on muscle size measures from ultrasound images are inadequate in precision. Radiomics has been widely used in various medical studies, but its validity for the evaluation of muscle atrophy has not been fully explored.

METHODS

This study presents a radiomics analysis for muscle atrophy evaluation using ultrasound images. The hindlimb unloading rat model was developed to simulate weightlessness muscle atrophy and ultrasound images of the hind limbs were acquired for both the hindlimb unloaded (HU) and control groups during a 21-day HU period. A total of 368 radiomics features were extracted and the stable and informative features were selected through a two-stage feature selection procedure. The feature change trajectory of the stable features was analyzed using the hierarchical clustering method. Finally, an adaptive longitudinal feature selection and grading network, ALNet, was developed to evaluate muscle atrophy.

RESULTS

The clustering trajectories of ultrasound image features showed similar trends to the changes in muscle atrophy at the molecular level. The best grading accuracy achieved by the ALNet was 79.5% for the Soleus (Sol) muscle and 82.6% for the Gastrocnemius (Gas) muscle.

CONCLUSION

The test-retest is essential in performing radiomics analysis on ultrasound images. The longitudinal feature selection is important for muscle atrophy grading. The ultrasound image features of the Gas muscle have better discrimination ability than that of the Sol muscle. This study proves for the first time the capability of ultrasound image features for muscle atrophy evaluation.

Evaluation of Feature Selection Methods for Mammographic Breast Cancer Diagnosis in a Unified Framework

Over recent years, feature selection (FS) has gained more attention in intelligent diagnosis. This study is aimed at evaluating FS methods in a unified framework for mammographic breast cancer diagnosis. After FS methods generated rank lists according to feature importance, the framework added features incrementally as the input of random forest which performed as the classifier for breast lesion classification. In this study, 10 FS methods were evaluated and the digital database for screening mammography (1104 benign and 980 malignant lesions) was analyzed. The classification performance was quantified with the area under the curve (AUC), and accuracy, sensitivity, and specificity were also considered. Experimental results suggested that both infinite latent FS method (AUC, 0.866 ± 0.028) and RELIEFF (AUC, 0.855 ± 0.020) achieved good prediction (AUC ≥ 0.85) when 6 features were used, followed by correlation-based FS method (AUC, 0.867 ± 0.023) using 7 features and WILCOXON (AUC, 0.887 ± 0.019) using 8 features. The reliability of the diagnosis models was also verified, indicating that correlation-based FS method was generally superior over other methods. Identification of discriminative features among high-throughput ones remains an unavoidable challenge in intelligent diagnosis, and extra efforts should be made toward accurate and efficient feature selection.

An ensemble of the iCluster method to analyze longitudinal lncRNA expression data for psoriasis patients

Background

Psoriasis is an immune-mediated, inflammatory disorder of the skin with chronic inflammation and hyper-proliferation of the epidermis. Since psoriasis has genetic components and the diseased tissue of psoriasis is very easily accessible, it is natural to use high-throughput technologies to characterize psoriasis and thus seek targeted therapies. Transcriptional profiles change correspondingly after an intervention. Unlike cross-sectional gene expression data, longitudinal gene expression data can capture the dynamic changes and thus facilitate causal inference.

Methods

Using the iCluster method as a building block, an ensemble method was proposed and applied to a longitudinal gene expression dataset for psoriasis, with the objective of identifying key lncRNAs that can discriminate the responders from the non-responders to two immune treatments of psoriasis.

Results

Using support vector machine models, the leave-one-out predictive accuracy of the 20-lncRNA signature identified by this ensemble was estimated as 80%, which outperforms several competing methods. Furthermore, pathway enrichment analysis was performed on the target mRNAs of the identified lncRNAs. Of the enriched GO terms or KEGG pathways, proteasome, and protein deubiquitination is included. The ubiquitination-proteasome system is regarded as a key player in psoriasis, and a proteasome inhibitor to target ubiquitination pathway holds promises for treating psoriasis.

Conclusions

An integrative method such as iCluster for multiple data integration can be adopted directly to analyze longitudinal gene expression data, which offers more promising options for longitudinal big data analysis. A comprehensive evaluation and validation of the resulting 20-lncRNA signature is highly desirable.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40246-021-00323-6.

GEE-TGDR: A Longitudinal Feature Selection Algorithm and Its Application to lncRNA Expression Profiles for Psoriasis Patients Treated with Immune Therapies

With the fast evolution of high-throughput technology, longitudinal gene expression experiments have become affordable and increasingly common in biomedical fields. Generalized estimating equation (GEE) approach is a widely used statistical method for the analysis of longitudinal data. Feature selection is imperative in longitudinal omics data analysis. Among a variety of existing feature selection methods, an embedded method-threshold gradient descent regularization (TGDR)-stands out due to its excellent characteristics. An alignment of GEE with TGDR is a promising area for the purpose of identifying relevant markers that can explain the dynamic changes of outcomes across time. We proposed a new novel feature selection algorithm for longitudinal outcomes-GEE-TGDR. In the GEE-TGDR method, the corresponding quasilikelihood function of a GEE model is the objective function to be optimized, and the optimization and feature selection are accomplished by the TGDR method. Long noncoding RNAs (lncRNAs) are posttranscriptional and epigenetic regulators and have lower expression levels and are more tissue-specific compared with protein-coding genes. So far, the implication of lncRNAs in psoriasis remains largely unexplored and poorly understood even though some evidence in the literature supports that lncRNAs and psoriasis are highly associated. In this study, we applied the GEE-TGDR method to a lncRNA expression dataset that examined the response of psoriasis patients to immune treatments. As a result, a list including 10 relevant lncRNAs was identified with a predictive accuracy of 70% that is superior to the accuracies achieved by two competitive methods and meaningful biological interpretation. A widespread application of the GEE-TGDR method in omics longitudinal data analysis is anticipated.

Incorporating Pathway Information into Feature Selection towards Better Performed Gene Signatures

To analyze gene expression data with sophisticated grouping structures and to extract hidden patterns from such data, feature selection is of critical importance. It is well known that genes do not function in isolation but rather work together within various metabolic, regulatory, and signaling pathways. If the biological knowledge contained within these pathways is taken into account, the resulting method is a pathway-based algorithm. Studies have demonstrated that a pathway-based method usually outperforms its gene-based counterpart in which no biological knowledge is considered. In this article, a pathway-based feature selection is firstly divided into three major categories, namely, pathway-level selection, bilevel selection, and pathway-guided gene selection. With bilevel selection methods being regarded as a special case of pathway-guided gene selection process, we discuss pathway-guided gene selection methods in detail and the importance of penalization in such methods. Last, we point out the potential utilizations of pathway-guided gene selection in one active research avenue, namely, to analyze longitudinal gene expression data. We believe this article provides valuable insights for computational biologists and biostatisticians so that they can make biology more computable.