Figure 1.
AML cell lines and patients with AML separate into 2 sphingolipidomic clusters that differ in their abundance of hexosylceramide and sphingomyelin. (A) Strategy to identify sphingolipidomic subtypes in AML. Sphingolipidomics of hexosylceramide (Hex), sphingomyelin (SM), ceramide (Cer), and LCB (comprised of sphingosine and its derivatives) was performed on primary AML samples and AML cell lines by liquid chromatography mass spectrometry, and the normalized data were consensus clustered to identify a stable number of sphingolipid clusters. Cluster-specific gene signatures were extracted to train a Hex-SM classifier that infers sphingolipidomic subtype from RNA sequencing (RNASeq). (B) Sphingolipidomic heterogeneity is similar in AML cell lines and patient samples but distinct from normal CD34+ bone marrow. Normalized sphingolipidomics for normal bone marrow samples (magenta, n = 6), primary AML samples (purple, n = 213), and AML cell lines (orange, n = 30) were displayed by Uniform Manifold Approximation and Projection (UMAP). (C) Row-standardized lipid abundances organized by the sphingolipid family: Hex, SM, Cer, and LCB. The HexlowSMhigh and HexhighSMlow consensus clusters are separately clustered and annotated as cell lines (orange) and patient samples (purple). (D-E) Normalized Z-scores of lipid species within the Hex (D) and SM (E) families were summed and differences between consensus clusters were assessed by the Mann-Whitney test with continuity correction. Colors indicate the sample type: AML cell lines (orange, n = 30) and primary samples (purple, n = 213).

AML cell lines and patients with AML separate into 2 sphingolipidomic clusters that differ in their abundance of hexosylceramide and sphingomyelin. (A) Strategy to identify sphingolipidomic subtypes in AML. Sphingolipidomics of hexosylceramide (Hex), sphingomyelin (SM), ceramide (Cer), and LCB (comprised of sphingosine and its derivatives) was performed on primary AML samples and AML cell lines by liquid chromatography mass spectrometry, and the normalized data were consensus clustered to identify a stable number of sphingolipid clusters. Cluster-specific gene signatures were extracted to train a Hex-SM classifier that infers sphingolipidomic subtype from RNA sequencing (RNASeq). (B) Sphingolipidomic heterogeneity is similar in AML cell lines and patient samples but distinct from normal CD34+ bone marrow. Normalized sphingolipidomics for normal bone marrow samples (magenta, n = 6), primary AML samples (purple, n = 213), and AML cell lines (orange, n = 30) were displayed by Uniform Manifold Approximation and Projection (UMAP). (C) Row-standardized lipid abundances organized by the sphingolipid family: Hex, SM, Cer, and LCB. The HexlowSMhigh and HexhighSMlow consensus clusters are separately clustered and annotated as cell lines (orange) and patient samples (purple). (D-E) Normalized Z-scores of lipid species within the Hex (D) and SM (E) families were summed and differences between consensus clusters were assessed by the Mann-Whitney test with continuity correction. Colors indicate the sample type: AML cell lines (orange, n = 30) and primary samples (purple, n = 213).

Close Modal

or Create an Account

Close Modal
Close Modal