LDAformer: predicting lncRNA-disease associations based on topological feature extraction and Transformer encoder

Abstract

The identification of long noncoding RNA (lncRNA)-disease associations is of great value for disease diagnosis and treatment, and it is now commonly used to predict potential lncRNA-disease associations with computational methods. However, the existing methods do not sufficiently extract key features during data processing, and the learning model parts are either less powerful or overly complex. Therefore, there is still potential to achieve better predictive performance by improving these two aspects. In this work, we propose a novel lncRNA-disease association prediction method LDAformer based on topological feature extraction and Transformer encoder. We construct the heterogeneous network by integrating the associations between lncRNAs, diseases and micro RNAs (miRNAs). Intra-class similarities and inter-class associations are presented as the lncRNA-disease-miRNA weighted adjacency matrix to unify semantics. Next, we design a topological feature extraction process to further obtain multi-hop topological pathway features latent in the adjacency matrix. Finally, to capture the interdependencies between heterogeneous pathways, a Transformer encoder based on the global self-attention mechanism is employed to predict lncRNA-disease associations. The efficient feature extraction and the intuitive and powerful learning model lead to ideal performance. The results of computational experiments on two datasets show that our method outperforms the state-of-the-art baseline methods. Additionally, case studies further indicate its capability to discover new associations accurately.

GCN-Based Heterogeneous Complex Feature Learning to Enhance Predictability for LncRNA–Disease Associations

“…On the other hand, Data set2, a newly constructed data set in 2022, exhibits a sparsity ratio of 1:55. While Data set1 is a well-established data set with extensive usage, its construction process relies on certain logical presuppositions as outlined in the original literature . In contrast, Data set2 was constructed respecting the original literature’s evidence records in Lnc2Cancer and LncRNADisease, without introducing any logical presuppositions in the process.…”

“…The performance of our HGCNLDA was evaluated on two benchmark data sets with collection and preprocession described in the literature 19 and literature, 30 respectively:…”

“…In 2022, Fan et al 29 proposed GCRFLDA, a novel prediction method that constructed an encoder with a conditional random field and attention mechanism to learn efficient embeddings of nodes, alongside a decoder layer to score LDAs. Also, in 2022, Zhou et al 30 proposed LDAformer, a novel LDA prediction model based on topological feature extraction and a transformer encoder. LDAformer designed a topological feature extraction process to capture multihop topological pathway features latent in the heterogeneous network and used a transformer encoder based on global self-attention to infer LDAs by capturing interdependencies between heterogeneous pathways.…”

Recent advances in machine learning methods for predicting LncRNA and disease associations

“…VADLP is used to extract, encode, and adaptively integrate a predictive model for multi-layer representation (Sheng et al, 2021). A three-level heterogeneous graph with inter-layer and intra-layer edges weighting is constructed to facilitate node attribute embedding and pairwise topology extraction for r a n d o m w a n d e r i n g, a n d t h e m o d e l d e fi n e s t h r e e representations, including node attributes, pairwise topology, (Zhou et al, 2022). This method constructs the heterogeneous network by integrating the associations between lncRNAs, diseases and micro RNAs (miRNAs).…”

“…Zhou et al. propose a novel lncRNA-disease association prediction method LDAformer based on topological feature extraction and Transformer encoder ( Zhou et al., 2022 ). This method constructs the heterogeneous network by integrating the associations between lncRNAs, diseases and micro RNAs (miRNAs).…”