Graph representation learning in biomedicine and healthcare

In the paper

“Further, essential genes are often found in hubs of a molecular network whereas non-essential genes (e.g., those associated with disease) are located on the periphery [13].”

What the sources say

• Partly supported

  The ‘Location of disease genes within networks’ section says essential genes tend to be associated with hubs, while non-essential disease genes tend toward the functional periphery. The citing sentence broadens that second finding to non-essential genes generally, which the passage does not establish. (LLM, conf=0.86)

  cites [13] Barabási, A.-L., Gulbahce, N. & Loscalzo, J. Network medicine: a network-based approach to human disease. Nat. Rev. Genet. 12, 56–68 (2011). sourceExperimental — accuracy not yet measured

In the paper

“Similar to proteins, small compounds are modeled as 2D and 3D molecular graphs such that nodes are atoms and edges are bonds.”

What the sources say

• Partly supported

  Section 5.1 (p. 4) describes MUTAG as a dataset of molecular graphs whose relations can indicate atomic bonds. Section 2 (p. 2) defines graph nodes and edges generally, but our copy does not say that small compounds are modeled as both 2D and 3D graphs or explicitly identify their nodes as atoms. (LLM, conf=0.85)

  cites [83] Schlichtkrull, M. et al. Modeling relational data with graph convolutional networks. In European Semantic Web Conference (eds Gangemi, A. et al.) 593–607 (Springer, 2018). sourceExperimental — accuracy not yet measured
• ?Could not verify

  The Methods section on Topological Data Analysis describes graphs whose nodes are clusters of drugs, targets, or diseases and whose edges indicate shared members. The Results section discusses drug structural similarity, but neither passage says small compounds are modeled as 2D or 3D molecular graphs with atoms as nodes and bonds as edges. Searches for those terms and for molecular representation found no supporting passage in our copy; page numbers were unavailable. (LLM, conf=0.76)

  cites [166] Wang, R., Li, S., Cheng, L., Wong, M. H. & Leung, K. S. Predicting associations among drugs, targets and diseases by tensor decomposition for drug repositioning. BMC Bioinf. 20, 628 (2019). sourceExperimental — accuracy not yet measured

In the paper

“Applied to such networks, graph representation learning can impute scRNA-seq data [156, 157], predict cell clusters [157, 158], etc.”

What the sources say

• ?Could not verify

  Section 15.7 describes using topological data analysis to group similar drug compounds and select leads. Searches for single-cell RNA sequencing, gene or cell networks, imputation, and cell-cluster prediction found no passage addressing those claims; the paper’s compound clusters do not establish a claim about cell clusters. (LLM, conf=0.76)

  cites [156] Alagappan, M., Jiang, D., Denko, N. & Koong, A. C. in Tumor Microenvironment (eds Koumenis, C. et al.) 253–268 (Springer, 2016). sourceExperimental — accuracy not yet measured
• ?Could not verify

  The Methods section (‘Graph embedding for feature learning’; page unavailable) describes node2vec embeddings of drug–target and drug/target similarity graphs for predicting drug–target interactions. Our copy does not address scRNA-seq imputation or cell-cluster prediction; searches for those terms and gene or cell similarity networks found no supporting passage. (LLM, conf=0.76)

  cites [157] Thafar, M. A. et al. DTiGEMS+: drug–target interaction prediction using graph embedding, graph mining, and similarity-based techniques. J. Cheminformatics 12, 44 (2020). sourceExperimental — accuracy not yet measured
• ?Could not verify

  The Materials and method section describes graph embedding on drug–drug and target–target similarity networks to predict drug–target interactions. Searches for scRNA-seq imputation and cell clustering found no passage addressing either claimed outcome. The paper’s application of graph embedding does not establish those single-cell applications. (LLM, conf=0.77)

  cites [158] Thafar, M. A. et al. DTi2Vec: Drug–target interaction prediction using network embedding and ensemble learning. J. Cheminformatics 13, 71 (2021). sourceExperimental — accuracy not yet measured

In the paper

“As a result, diagnosing diseases requires knowledge about alterations in the transcription of coding genes to capture genome-wide associations driving disease acquisition and progression.”

In the paper

“Graph representation learning is challenging because graphs contain complex topographical structure, have no fixed node ordering or reference points, and are comprised of many different kinds of entities (nodes) and various types of interactions (edges) relating them to each other.”

In the paper

“We also capture the breadth of ways in which representation learning has dramatically improved the state-of-the-art in biomedical machine learning.”

In the paper

“Biomedical networks (or graphs) are universal descriptors for systems of interacting elements, from molecular interactions and disease co-morbidity to healthcare systems and scientific knowledge.”

In the paper

“These approaches are transductive embedding methods where the encoder function is a simple embedding lookup (Figure 2).”

In the paper

“Graph neural networks (GNNs) are a class of neural networks designed for graph-structured datasets (Figure 2).”

In the paper

“Such functions are optimized to embed the input network so that nodes with similar network neighborhoods are embedded close together in the embedding space, and algebraic operations performed in this learned space reflect the network’s topology.”

In the paper

“we put forward an observation that long-standing principles of systems biology and medicine—while often unspoken in machine learning research—provide the conceptual grounding for representation learning on graphs, explain its current successes and limitations, and even inform future advancements.”

In the paper

“Graph representation learning methods allow us to analyze heterogeneous networks of multimodal data and make predictions across domains, from genomic level data (e.g., gene expression, copy number information) to clinically relevant data (e.g., pathophysiology, tissue localization).”

In the paper

“Its core idea is to learn how to represent nodes (or larger graph structures) in a network as points in a low-dimensional space, where the geometry of this space is optimized to reflect the structure of interactions between nodes.”

In the paper

“Comparing transcriptomic profiles from healthy individuals to those of patients with a specific disease informs clinicians of its causal genes.”

In the paper

“We synthesize a spectrum of algorithmic approaches that, at their core, leverage graph topology to embed networks into compact vector spaces.”

In the paper

“GNNs (as described in Section 4.1) are commonly used to generate representations of proteins based on chemical (e.g., free electrons’ and protons donors’ locations) and geometric (e.g., distance-dependent curvature) features to predict protein pocket-ligand and protein-protein interactions sites [116]; intra- and inter-molecule residue contact graphs to predict intra- and inter-molecular energies,”

What the sources say

• Not checked yetcites [116] Li, D. & Gao, J. Towards perturbation prediction of biological networks using deep learning. Sci. Rep. 9, 11941 (2019). source
• Not checked yetcites [117] Liu, Y., Yuan, H., Cai, L., & Ji, S. Deep learning of high-order interactions for protein interface prediction. In Proc. ACM SIGKDD Conference on Knowledge Discovery and Data Mining 679–687 (ACM, 2020). source
• Not checked yetcites [111] Elton, D. C., Boukouvalas, Z., Fuge, M. D. & Chung, P. W. Deep learning for molecular design—a review of the state of the art. Mol. Syst. Des. Eng. 4, 828–849 (2019). source

In the paper

“Further, injecting gene expression data into PPI networks has identified disease biomarkers, which are then used to more accurately classify diseases of interest.”

In the paper

“They learn compact representations of graph elements, their attributes, and supervised labels if any.”

In the paper

“Akin to how deep learning on images and sequences has revolutionized image analysis and natural language processing, graph representation learning is poised to transform the study of complex systems.”

In the paper

“Shallow embedding methods optimize a compact vector space such that points that are close in the graph are mapped to nearby points in the embedding space, which is measured by a predefined distance function or an outer product.”

In the paper

“Deep generative models address the challenge by estimating distributional graph properties based on a dataset of graphs 𝒢 and inferring graph structures using such optimized distributions.”

In the paper

“Additionally, GNNs can optimize supervised signals and graph structure simultaneously, whereas shallow embedding approaches require a two-stage approach to achieve the same.”

In the paper

“Since protein structures are folded into complex 3D structures, it is natural to represent them as graphs.”

In the paper

“Gene Ontology (GO) terms [125] are a standardized vocabulary for describing molecular functions, biological processes, and cellular locations of gene products [126].”

What the sources say

• Not checked yetcites [125] Hasibi, R. & Michoel, T. A Graph feature auto-encoder for the prediction of unobserved node features on biological networks. BMC Bioinform 22, 525 (2021). source
• Not checked yetcites [126] Cao, M. et al. Going the distance for protein function prediction: a new distance metric for protein interaction networks. PLoS ONE 8, e76339 (2013). source

In the paper

“Learning the topology of graphs containing drug, protein, and disease nodes has improved predictions of candidate drugs for treating a disease, identification of potential off-target effects, and prioritization of novel drug combinations.”

What the sources say

• Not checked yetcites [103] Fan, K. & Zhang, Y. Pseudo2GO: A graph-based deep learning method for pseudogene function prediction by borrowing information from coding genes. Front. Genet. 11, 807 (2020). source

In the paper

“Hence, artificial intelligence methods that produce representations can be thought of as differentiable engines of key network biomedicine principles.”

In the paper

“Diseases are classified based on the presenting symptoms of patients, which can be caused by molecular dysfunctions, such as genetic mutations.”

In the paper

“More concretely, representation learning specifies deep, non-linear transformation functions that map nodes to points in a compact vector space, termed embeddings.”

In the paper

“As biomedical networks can be multimodal and massive, special consideration is needed to scale GNNs to large and heterogeneous networks.”

In the paper

“They have been built as a hierarchical graph that GNNs then leverage to learn dependencies of the terms [126], or directly used as protein function labels [106, 127].”

What the sources say

• Not checked yetcites [126] Cao, M. et al. Going the distance for protein function prediction: a new distance metric for protein interaction networks. PLoS ONE 8, e76339 (2013). source
• Not checked yetcites [106] Hamilton, W. L., Ying, Z. & Leskovec, J. Inductive representation learning on large graphs. In Proc. International Conference on Neural Information Processing Systems (eds Guyon, I. et al.) (Curran Associates, 2017).
• Not checked yetcites [127] Dey, T. K. & Mandal, S. Protein classification with improved topological data analysis. In 18th International Workshop on Algorithms in Bioinformatics (WABI) (eds Parida, L. & Ukkonen, E.) 6:1–6:13 (Schloss Dagstuhl–Leibniz-Zentrum fuer Informatik, 2018).

In the paper

“To this end, graph representation learning methods, such as GNNs, have been developed to learn robust and meaningful representations of molecules despite the incomplete interactome [102] and to inductively infer new edges between pairs of nodes [154].”

What the sources say

• Not checked yetcites [102] Yin, N. et al. Synergistic and antagonistic drug combinations depend on network topology. PLoS ONE 9, e93960 (2014). source
• Not checked yetcites [154] Coley, C. W. et al. A graph-convolutional neural network model for the prediction of chemical reactivity. Chem. Sci. 10, 370–377 (2019). source

In the paper

“Spatial molecular profiling can measure both gene expression at the cellular level and location of cells in a tissue [160].”

What the sources say

• Not checked yetcites [160] Jiang, M. et al. Drug–target affinity prediction using graph neural network and contact maps. RSC Adv 10, 20701–20712 (2020). source

In the paper

“Edges can also be added to indicate pairwise spatial distance between two atoms [72], or directed with information on bond angles and rotations incorporated into the molecular graph [84].”

What the sources say

• Not checked yetcites [72] Hu, Z., Dong, Y., Wang, K. & Sun, Y. Heterogeneous graph transformer. In Proc. ACM Web Conference 2704–2710 (ACM, 2020). source
• Not checked yetcites [84] Wang, X. et al. Heterogeneous graph attention network. In Proc. ACM Web Conference (eds Liu, L. & White, R.) 2022–2032 (ACM, 2019). source

In the paper

“Creating cell neighborhood and spatial gene expression graphs require a distance metric, as edges are determined based on spatial proximity, while gene co-expression and molecular similarity graphs need a threshold applied on the gene expression data [35].”

What the sources say

• Not checked yetcites [35] Hetzel, L., Fischer, D. S., Günnemann, S. & Theis, F. J. Graph representation learning for single cell biology. Curr. Opin. Syst. Biol. 28, 100347 (2021). source

In the paper

“As a result, spatial transcriptomics data can be used to construct cell graphs [161], spatial gene expression graphs [162], gene co-expression networks, or molecular similarity graphs [35].”

What the sources say

• Not checked yetcites [161] Chen, L. et al. TransformerCPI: improving compound–protein interaction prediction by sequence-based deep learning with self-attention mechanism and label reversal experiments. Bioinformatics 36, 4406–4414 (2020). source
• Not checked yetcites [162] Quan, Z., Guo, Y., Lin, X., Wang, Z.-J. & Zeng, X. GraphCPI: Graph neural representation learning for compound-protein interaction. In BIBM 717–722 (2019). source
• Not checked yetcites [35] Hetzel, L., Fischer, D. S., Günnemann, S. & Theis, F. J. Graph representation learning for single cell biology. Curr. Opin. Syst. Biol. 28, 100347 (2021). source

In the paper

“Modern drug discovery requires elucidating a candidate drug’s chemical structure, identifying its drug targets, quantifying its efficacy and toxicity, and detecting its potential side effects [13, 14, 32, 164].”

What the sources say

• Not checked yetcites [13] Barabási, A.-L., Gulbahce, N. & Loscalzo, J. Network medicine: a network-based approach to human disease. Nat. Rev. Genet. 12, 56–68 (2011). source
• Not checked yetcites [14] Hu, J. X., Thomas, C. E. & Brunak, S. Network biology concepts in complex disease comorbidities. Nat. Rev. Genet. 17, 615–629 (2016). source
• Not checked yetcites [32] Rai, A., Shinde, P. & Jalan, S. Network spectra for drug-target identification in complex diseases: new guns against old foes. Appl. Netw. Sci. 3, 51 (2018). source
• Not checked yetcites [164] Lin, X. et al. DeepGS: Deep representation learning of graphs and sequences for drug-target binding affinity prediction. ECAI (2020).

In the paper

“Each atom and bond may include features, such as atomic mass, atomic number, and bond type, to be included in the model [83, 166].”

What the sources say

• Not checked yetcites [83] Schlichtkrull, M. et al. Modeling relational data with graph convolutional networks. In European Semantic Web Conference (eds Gangemi, A. et al.) 593–607 (Springer, 2018). source
• Not checked yetcites [166] Wang, R., Li, S., Cheng, L., Wong, M. H. & Leung, K. S. Predicting associations among drugs, targets and diseases by tensor decomposition for drug repositioning. BMC Bioinf. 20, 628 (2019). source

In the paper

“Also, two separate GNNs may be used to learn representations of protein and small molecule graphs for predicting drug-target affinity [173].”

What the sources say

• Not checked yetcites [173] Choi, E., Bahadori, M. T., Song, L., Stewart, W. F. & Sun, J. GRAM: graph-based attention model for healthcare representation learning. In Proc. 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining 787–795 (ACM, 2017). source

In the paper

“Medical images of patients, including histopathology slides, enable clinicians to comprehensively observe the effects of a disease on the patient’s body [188].”

What the sources say

• Not checked yetcites [188] Wu, T. et al. Representation learning of EHR data via graph-based medical entity embedding. In Advances in Neural Information Processing Systems Graph Representation Learning Workshop (2019).

In the paper

“For instance, Medical Subject Headings (MeSH) terms may be used to construct a drug-disease graph, from which latent representations of drugs and diseases are learned using various graph embedding algorithms, including DeepWalk and LINE (Section 2.1) [178].”

What the sources say

• Not checked yetcites [178] Zhou, Y. et al. CGC-Net: cell graph convolutional network for grading of colorectal cancer histology images. In IEEE International Conference on Computer Vision Workshops (IEEE, 2019). source

In the paper

“Concretely, TDA (refer to SI Note 3) transforms experimental data into a graph where nodes represent compounds and edges indicate a level of similarity between them [169].”

What the sources say

• Not checked yetcites [169] Alsentzer, E., Finlayson, S. G., Li, M. M. & Zitnik, M. Subgraph neural networks. In Proc. International Conference on Neural Information Processing Systems (eds Larochelle, H. et al.) 8017–8029 (Curran Associates, 2020).

In the paper

“Graph representation learning methods enables in silico modeling of drug action, which allows for more efficient ranking of candidate drugs for repurposing, such as by considering gene expression data, gene ontologies, drug similarity, and other clinically relevant data regarding side effects and indications.”

In the paper

“K-nearest neighbors, for instance, is a common used method to construct drug and target similarity networks [171].”

What the sources say

• Not checked yetcites [171] Barisoni, L., Lafata, K. J., Hewitt, S. M., Madabhushi, A. & Balis, U. G. Digital pathology and computational image analysis in nephropathology. Nat. Rev. Nephrol. 16, 669–685 (2020). source

In the paper

“Further, cell morphology and interactions, tissue morphology and spatial distribution, cell-to-tissue hierarchies, and spatial distribution of cells with respect to tissues can be captured in a cell-to-tissue graph, upon which a hierarchical GNN can learn representations using these different data modalities [189].”

What the sources say

• Not checked yetcites [189] Mao, C., Yao, L. & Luo, Y. MedGCN: medication recommendation and lab test imputation via graph convolutional networks. J. Biomed. Inform. 127, 104000 (2022). source

In the paper

“Intuitively, message passing steps (i.e., in GNNs (Section 2.2)) aggregate information from neighboring atoms and bonds to learn the local chemistry of each atom [166].”

What the sources say

• Not checked yetcites [166] Wang, R., Li, S., Cheng, L., Wong, M. H. & Leung, K. S. Predicting associations among drugs, targets and diseases by tensor decomposition for drug repositioning. BMC Bioinf. 20, 628 (2019). source

In the paper

“Part of the drug discovery pipeline is minimizing adverse drug events [33, 34].”

What the sources say

• Not checked yetcites [33] David, L., Thakkar, A., Mercado, R. & Engkvist, O. Molecular representations in AI-driven drug discovery: a review and practical guide. J. Cheminformatics 12, 56 (2020). source
• Not checked yetcites [34] Wieder, O. et al. A compact review of molecular property prediction with graph neural networks. Drug Discov. Today. Technol. 37, 1–12 (2020). source

In the paper

“EHRs are often supplemented with other modalities, such as diseases, symptoms, molecular data, drug interactions, etc [7, 206, 211, 212].”

What the sources say

• Not checked yetcites [7] Nelson, C. A., Butte, A. J. & Baranzini, S. E. Integrating biomedical research and electronic health records to create knowledge-based biologically meaningful machine-readable embeddings. Nat. Commun. 10, 3045 (2019). source
• Not checked yetcites [206] Lanchantin, J. & Qi, Y. Graph convolutional networks for epigenetic state prediction using both sequence and 3D genome data. Bioinformatics 36, i659–i667 (2020). source
• Not checked yetcites [211] Ying, Z., Bourgeois, D., You, J., Zitnik, M. & Leskovec, J. GNNExplainer: generating explanations for graph neural networks. In Proc. International Conference on Neural Information Processing Systems (ed Wallach, H. et al.) (Curran Associates, 2019).
• Not checked yetcites [212] Agarwal, C., Lakkaraju, H. & Zitnik, M. Towards a unified framework for fair and stable graph representation learning. In Proc. Machine Learning Research (eds de Campos, C. & Maathuis, M. H.) 2114–2124 (PMLR, 2021).

In the paper

“Alternatively, a patient graph may be constructed based on the similarity of patients, and patient embeddings learned by an LSTM-GNN architecture are optimized to predict patient outcomes [210].”

What the sources say

• Not checked yetcites [210] Machens, A. et al. An infectious disease model on empirical networks of human contact: bridging the gap between dynamic network data and contact matrices. BMC Infect. Dis. 13, 185 (2013). source

In the paper

“An ST-GCN [78] is designed to utilize the underlying spatial and temporal dependencies of EHR data for generating patient diagnoses [187].”

What the sources say

• Not checked yetcites [187] Rotmensch, M., Halpern, Y., Tlimat, A., Horng, S. & Sontag, D. Learning a health knowledge graph from electronic medical records. Sci. Rep. 7, 5994 (2017). source

In the paper

“Alternatively, TDA can generate graphs of whole-slide images, which include tissues from various patient sources (SI Note 3), and GNNs are then used to classify the stage of colon cancer [199].”

What the sources say

• Not checked yetcites [199] Li, L. et al. A method to learn embedding of a probabilistic medical knowledge graph: algorithm development. JMIR Med. Inf. 8, e17645 (2020). source

In the paper

“Meta-paths may alternatively be exploited in an EHR-derived KG to leverage higher order, semantically important relations for disease classification [214].”

What the sources say

• Not checked yetcites [214] Obermeyer, Z., Powers, B., Vogeli, C. & Mullainathan, S. Dissecting racial bias in an algorithm used to manage the health of populations. Science 366, 447–453 (2019). source

In the paper

“Node features for drugs and diseases can be initialized using Skipgram and then a GNN leveraging multi-layer message-passing can be applied to predict adverse drug events [211].”

What the sources say

• Not checked yetcites [211] Ying, Z., Bourgeois, D., You, J., Zitnik, M. & Leskovec, J. GNNExplainer: generating explanations for graph neural networks. In Proc. International Conference on Neural Information Processing Systems (ed Wallach, H. et al.) (Curran Associates, 2019).

In the paper

“In public health, spatial and temporal networks can model space- and time-dependent observations (e.g., disease states, susceptibility to infection [225]) to spot trends, detect anomalies, and interpret temporal dynamics.”

What the sources say

• Not checked yetcites [225] Tran, D. et al. Fast and precise single-cell data analysis using a hierarchical autoencoder. Nat. Commun. 12, 1029 (2021). source

In the paper

“Since multimodal data enables more robust predictions, GNNs (Section 2.2) have been applied to generate cell spatial graphs from histopathology images and then fuse genomic and transcriptomic data for predicting treatment response and resistance, histopathology grading, and patient survival [8].”

What the sources say

• Not checked yetcites [8] Chen, R. J. et al. Pathomic fusion: an integrated framework for fusing histopathology and genomic features for cancer diagnosis and prognosis. In IEEE Transactions on Medical Imaging Vol. 41, 757–770 (IEEE, 2022). source

In the paper

“Alternatively, attention mechanisms may be applied on EHR data and medical ontologies to capture the parent-child relationships [186, 204, 205].”

What the sources say

• Not checked yetcites [186] Hu, J. et al. SpaGCN: integrating gene expression, spatial location and histology to identify spatial domains and spatially variable genes by graph convolutional network. Nat. Methods 18, 1342–1351 (2021). source
• Not checked yetcites [204] Wang, T., Peng, Q., Liu, B., Liu, Y. & Wang, Y. Disease module identification based on representation learning of complex networks integrated from GWAS, eQTL summaries, and human interactome. Front. Bioeng. Biotechnol. 8, 418 (2020). source
• Not checked yetcites [205] Dekker, J. & Misteli, T. Long-range chromatin interactions. Cold Spring Harb. Perspect. Biol. 7, a019356 (2015). source

In the paper

“Additionally, because graphs can model long-range dependencies or interactions, we can model chromatin elements and the effects of their binding to regions across the genome as a network [220, 221].”

What the sources say

• Not checked yetcites [220] Stark, C. et al. BioGRID: a general repository for interaction datasets. Nucleic Acids Res 34, D535–D539 (2006). source
• Not checked yetcites [221] Efremova, M., Vento-Tormo, M., Teichmann, S. A. & Vento-Tormo, R. Cellphonedb: inferring cell–cell communication from combined expression of multi-subunit ligand–receptor complexes. Nat. Protoc. 15, 1484–1506 (2020). source

In the paper

“A probabilistic KG of EHR data, which include medical history, drug prescriptions, and laboratory examination results, has been used to consider the semantic relations between EHR entities in a shallow network embedding method (Section 2.1) [213].”

What the sources say

• Not checked yetcites [213] Zhang, X. & Zitnik, M. GNNGuard: defending graph neural networks against adversarial attacks. In Proc. International Conference on Neural Information Processing Systems (eds Larochelle, H. et al.) 9263–9275 (Curran Associates, 2020).