Disentangling non-specific and condition-specific transcriptomic responses: A framework from heat stress in yeast

Abstract

A central challenge in biology is to distinguish transcriptional responses that recur across heterogeneous stress conditions from those that are specific to particular experimental contexts. Here, using Saccharomyces cerevisiae as a model system, we present an integrative transcriptomic framework to address this issue from bulk population-level datasets. We combined 13 carefully curated public transcriptomes with two in-house datasets generated under controlled heat ramp and heat shock conditions. By integrating co-expression network analysis, functional enrichment, and promoter motif discovery, we identified gene modules associated either with recurrent non-specific stress responses or with condition-specific responses linked to gradual heating.

Because bulk transcriptomic measurements capture an integrated signal across cells that may differ in physiological status, including viable, injured, and non-cultivable subpopulations, our conclusions are intentionally framed at the population level rather than at the level of individual cells. Within this framework, heat ramp, in contrast to sudden heat shock, was associated with a broader and more structured population-level transcriptional program enriched in stress-response genes and regulatory motifs linked to increased survival.

Differential transcriptomic responses were associated with ribosome biogenesis, RNA processing, mitochondrial function, and proteostasis, while also showing that stress-responsive gene clusters displayed limited overlap with evolutionarily young genes. To summarize these findings, we propose conceptual models of the transcriptional organizations associated with heat ramp and heat shock and highlight their dependence on heating kinetics. Although developed in yeast, this framework provides a broadly applicable strategy for distinguishing recurrent from condition-specific transcriptomic patterns in heterogeneous stressed populations.