SMALF: miRNA-disease associations prediction based on stacked autoencoder and XGBoost

Abstract

Background Identifying miRNA and disease associations helps us understand disease mechanisms of action from the molecular level. However, it is usually blind, time-consuming, and small-scale based on biological experiments. Hence, developing computational methods to predict unknown miRNA and disease associations is becoming increasingly important. Results In this work, we develop a computational framework called SMALF to predict unknown miRNA-disease associations. SMALF first utilizes a stacked autoencoder to learn miRNA latent feature and disease latent feature from the original miRNA-disease association matrix. Then, SMALF obtains the feature vector of representing miRNA-disease by integrating miRNA functional similarity, miRNA latent feature, disease semantic similarity, and disease latent feature. Finally, XGBoost is utilized to predict unknown miRNA-disease associations. We implement cross-validation experiments. Compared with other state-of-the-art methods, SAMLF achieved the best AUC value. We also construct three case studies, including hepatocellular carcinoma, colon cancer, and breast cancer. The results show that 10, 10, and 9 out of the top ten predicted miRNAs are verified in MNDR v3.0 or miRCancer, respectively. Conclusion The comprehensive experimental results demonstrate that SMALF is effective in identifying unknown miRNA-disease associations.

Learning a confidence score and the latent space of a new Supervised Autoencoder for diagnosis and prognosis in clinical metabolomic studies

“…• Unlike classical autoencoders that assume a latent space modeling [23,24], we have used a supervised autoencoder (SAE). The real distributions of many datasets, including metabolomics datasets, are far from multi-gaussian mixtures.…”

“…In order to cope with this issue, some recent papers have proposed latent spaces with more complex distributions (e.g. mixtures of Gaussians [22,23,24]) on the latent vectors, but they are non-adaptive and unfortunately may not match the specific data distribution. In this work, we relaxed the parametric distribution assumption on the latent space to learn a non-parametric data distribution of clusters [25].…”

Identification of miRNA Signature Associated With Erectile Dysfunction in Type 2 Diabetes Mellitus by Support Vector Machine-Recursive Feature Elimination

“…Changes in their expression levels can affect multiple cellular processes and are used as molecular markers for diagnosis and follow-up ( Han et al, 2021 ). It is widely involved in pathological processes such as cancer ( Rupaimoole and Slack, 2017 ; Liu et al, 2021a ; Lei and Shu-Lin, 2021 ; Sheng et al, 2021 ; Tang et al, 2021 ), DM ( Vasu et al, 2019 ), cardiovascular events ( Barwari et al, 2016 ), and ED ( Ding et al, 2017 ). However, there are few related studies in DMED.…”

GCAEMDA: Predicting miRNA-disease associations via graph convolutional autoencoder

“…They integrated the prediction scores from different trained variational autoencoder models to infer unverified miRNA-disease associations. Liu et al [ 38 ] developed a computational model called to infer unknown miRNA-disease associations. SMALF first utilized the stacked autoencoder to extract miRNA and disease latent features from the original association matrix of miRNA-disease.…”