search

YosefLab / Compass

In-Silico Modeling of Metabolic Heterogeneity using Single-Cell Transcriptomes
BSD 3-Clause "New" or "Revised" License
96 stars 27 forks source link

Under pathological conditions, is it normal that all metabolic reactions are upregulated? #114

Open jouska540 opened 3 weeks ago

jouska540 commented 3 weeks ago

I am using compass software. I formed metacells from single-cell data using supercell and used it as the input for compass. As a result, I found that under pathological conditions, almost all metabolic reactions are upregulated. I am wondering if this is normal. image

xin0107 commented 3 weeks ago

I encountered a similar situation! As shown in the figure, my experimental design has a total of 4 experimental groups and one control group, NM. When compared to the control group, only the fourth experimental group showed significant differences in reaction activity. However, I also observed that the overall metabolic reactions of the different experimental groups compared to the control group were consistently either higher or lower. I am also suspicious that this might not be normal. Could you please help explain this? Additionally, the direction of metabolic reaction activity compared to the control group seems to be opposite to the total metabolic activity, which makes the situation even more peculiar to me. temp

jouska540 commented 3 weeks ago

image Hello. I have consulted the article "The glycolytic reaction PGAM unexpectedly restrains Th17 pathogenicity and Th17-dependent autoimmunity" recently published by the author's research group on bioRxiv. In Figure 4B, it can be seen that under the same glucose concentration, the metabolic activity of Th17n and Th17p is similar. Under different glucose concentration conditions, there are obvious differences in the overall metabolic activity of each cell. However, the author does not provide the differential changes of each metabolic reaction under different glucose concentrations. The article says that the compass hypothesis is established under the condition of sufficient metabolic raw materials. Then, under low glucose concentration, it actually does not conform to the compass hypothesis, so it is not easy to directly compare their metabolic reaction activities. In the original text, compass is only a comparison between cells with only differences in induction factors, that is, Th17n and Th17p. Perhaps for compass analysis, a condition with abundant metabolic substrates should be selected first to compare the correlation between metabolism and clinical characteristics (such as pathogenicity) of different subgroups of the same cell under this condition? Do you think this is possible?

I encountered a similar situation! As shown in the figure, my experimental design has a total of 4 experimental groups and one control group, NM. When compared to the control group, only the fourth experimental group showed significant differences in reaction activity. However, I also observed that the overall metabolic reactions of the different experimental groups compared to the control group were consistently either higher or lower. I am also suspicious that this might not be normal. Could you please help explain this? Additionally, the direction of metabolic reaction activity compared to the control group seems to be opposite to the total metabolic activity, which makes the situation even more peculiar to me. temp

xin0107 commented 3 weeks ago

Wow, I think you are right! I used bulk transcriptome data from mice under different experimental conditions, and the levels of metabolic substrates in their bodies may vary, so it might not be appropriate to directly use COMPASS for comparisons. Therefore, I should consider measuring the differences in nutrient concentrations and metabolic enzyme activities at the tissue level to reflect the differences in metabolic reactions. Thank you for your reply!