PT  - JOURNAL ARTICLE
AU  - Heather Jackson
AU  - Stephanie Menikou
AU  - Shea Hamilton
AU  - Andrew McArdle
AU  - Chisato Shimizu
AU  - Rachel Galassini
AU  - Honglei Huang
AU  - Jihoon Kim
AU  - Adriana Tremoulet
AU  - Marien de Jonge
AU  - Taco Kuijpers
AU  - Victoria Wright
AU  - Jane Burns
AU  - Climent Casals-Pascual
AU  - Jethro Herberg
AU  - Mike Levin
AU  - Myrsini Kaforou
ED  - ,
TI  - Kawasaki Disease patient stratification and pathway analysis based on host transcriptomic and proteomic profiles
AID  - 10.1101/2021.03.18.435948
DP  - 2021 Jan 01
TA  - bioRxiv
PG  - 2021.03.18.435948
4099  - http://biorxiv.org/content/early/2021/03/19/2021.03.18.435948.short
4100  - http://biorxiv.org/content/early/2021/03/19/2021.03.18.435948.full
AB  - The aetiology of Kawasaki Disease (KD), an acute inflammatory disorder of childhood, remains unknown despite various triggers of KD having been proposed. Host ‘omic profiles offer insights into the host response to infection and inflammation, with the interrogation of multiple ‘omic levels in parallel providing a more comprehensive picture. We used differential abundance analysis, pathway analysis, clustering and classification techniques to explore whether the host response in KD is more similar to the response to bacterial or viral infection at the transcriptomic and proteomic levels through comparison of ‘omic profiles from children with KD to those with bacterial and viral infections. Pathways activated in patients with KD included those involved in anti-viral and anti-bacterial responses. Unsupervised clustering showed that the majority of KD patients clustered with bacterial patients on both ‘omic levels, whilst application of diagnostic signatures specific for bacterial and viral infections revealed that many transcriptomic KD samples had low probabilities of having bacterial or viral infections, suggesting that KD may be triggered by a different process not typical of either common bacterial or viral infections. Clustering based on the transcriptomic and proteomic responses during KD revealed three clusters of KD patients on both ‘omic levels, suggesting heterogeneity within the inflammatory response during KD. The observed heterogeneity may reflect differences in the host response to a common trigger, or variation dependent on different triggers of the condition.Competing Interest StatementThe authors have declared no competing interest.CAACoronary artery aneurysmDBDefinite bacterialDRSDisease risk scoreDVDefinite viralFDRFalse discovery rateHCHealthy control KD Kawasaki DiseaseLFCLog-fold changeSDASignificantly differentially abundant