# Analysis of Drug repurposing Knowledge graphs for Covid-19

arXiv: https://arxiv.org/abs/2212.03911 · HTML: https://mlsanity.com/paper/2212.03911/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 9 of its 14 references that are in the map.
- First posted on arXiv 7 Dec 2022.

## Builds on (most cited first, 9 of 9)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Embedding Entities and Relations for Learning and Inference in Knowledge Bases](https://mlsanity.com/paper/1412.6575/) (2014)
- [Complex Embeddings for Simple Link Prediction](https://mlsanity.com/paper/1606.06357/) (2016)
- [RotatE: Knowledge Graph Embedding by Relational Rotation in Complex Space](https://mlsanity.com/paper/1902.10197/) (2019)
- [A Survey on Knowledge Graphs: Representation, Acquisition and Applications](https://mlsanity.com/paper/2002.00388/) (2020)
- [Composition-based Multi-Relational Graph Convolutional Networks](https://mlsanity.com/paper/1911.03082/) (2019)
- [HittER: Hierarchical Transformers for Knowledge Graph Embeddings](https://mlsanity.com/paper/2008.12813/) (2020)
- [Drug Repurposing for COVID-19 via Knowledge Graph Completion](https://mlsanity.com/paper/2010.09600/) (2020)
- [Few-shot link prediction via graph neural networks for Covid-19 drug-repurposing](https://mlsanity.com/paper/2007.10261/) (2020)

## Built on by (most cited first, 1 of 1)

- [MEDAKA: Construction of Biomedical Knowledge Graphs Using Large Language Models](https://mlsanity.com/paper/2509.26128/) (2025)

## Abstract (arXiv)

Knowledge graph (KG) is used to represent data in terms of entities and structural relations between the entities. This representation can be used to solve complex problems such as recommendation systems and question answering. In this study, a set of candidate drugs for COVID-19 are proposed by using Drug repurposing knowledge graph (DRKG). DRKG is a biological knowledge graph constructed using a vast amount of open source biomedical knowledge to understand the mechanism of compounds and the related biological functions. Node and relation embeddings are learned using knowledge graph embedding models and neural network and attention related models. Different models are used to get the node embedding by changing the objective of the model. These embeddings are later used to predict if a candidate drug is effective to treat a disease or how likely it is for a drug to bind to a protein associated to a disease which can be modelled as a link prediction task between two nodes. RESCAL performed the best on the test dataset in terms of MR, MRR and Hits@3.