Thank you for building the model. I would like to know if it's possible for the reaction penalty output to be completely similar between samples. I'm applying the algorithm to bulk RNA-seq data and the number of samples is not large (<100), which may explain the results. I don't expect the metabolic profile of the samples to be different due to the treatment being used (sleep deprivation, which should significantly perturb the metabolome) and the counts are quite different between samples, again making the output unexpected. You mentioned in another closed thread that statistical significance is a concern for bulk RNA-seq due to potentially small sample sizes; could the algorithm also perform less well by itself when processing few samples? Thanks a lot for any guidance.
Thank you for building the model. I would like to know if it's possible for the reaction penalty output to be completely similar between samples. I'm applying the algorithm to bulk RNA-seq data and the number of samples is not large (<100), which may explain the results. I don't expect the metabolic profile of the samples to be different due to the treatment being used (sleep deprivation, which should significantly perturb the metabolome) and the counts are quite different between samples, again making the output unexpected. You mentioned in another closed thread that statistical significance is a concern for bulk RNA-seq due to potentially small sample sizes; could the algorithm also perform less well by itself when processing few samples? Thanks a lot for any guidance.