Fig. 3: Single-cell quantitative analysis of lipogenic acetyl-CoA production and heterogeneity.

a, Comparison of bulk (top) and single-cell (bottom) analysis of isotopologue distribution for palmitate after 72 h of ACLY gene silencing and culture in the presence of ¹³C₆-glucose. Bulk data, generated by saponification and subsequent LC–MS, are displayed as mean ± s.d. across three replicates. For single-cell data, generated by MALDI with AIF, black lines show average values. b, Normalized single-cell isotopologue distributions for two individual cells, one from the control and the other from ACLYkd oligo 1 (shown as bar plots). Lines show fit of the fatty acid labelling binomial model: horizontal lines for M + 0 showing an estimated uptake and connected lines for M + 2i showing (1 − uptake) × binomial(i), where i is a variable from 1 to 8 indicating the even isotopic peaks. Legend shows parameters of the binomial model fit. c, Single-cell analysis of the estimated acetyl-CoA pool labelling degree (p) as calculated using the fatty acid labelling binominal model for the three conditions: control (grey), ACLYkd oligo 1 (blue) and ACLYkd oligo 2 (orange). Green dashed line shows 95% quantile of the GFP intensity distribution in the control condition, which was used to classify cells as GFP^pos versus GFP^neg. d, Spatial metabolic imaging of de novo fatty synthesis for the control, ACLYkd oligo 1 and ACLYkd oligo 2 conditions. Abundance of different palmitate isotopologue peaks is displayed in different channels: M + 0 (blue), M + 2 (green) and M + 8 (red). Each channel is normalized to the TIC.