Biochemical processes such as metabolism involve networks of thousands of reactions and are characterized by complex topologies. Energy conversion is central in these processes, and yet assessing it is problematic precisely due to the scale and topology of associated networks. I will describe how a general theory of nonequilibrium thermodynamics of chemical reaction networks enables a systematic characterization and assessment of such conversion. In the second part of my presentation, I will use some of the tools developed in thermodynamics and statistical mechanics for elucidating the notion of force in a seemingly unrelated class of processes: evolutionary population dynamics. I will clarify how one of these forces explains the intrinsic emergence of robustness of organismal reproduction to mutations without explicit selection for this trait. In the last part of my presentation, I will turn to ecological population dynamics involving microbes. Metabolic processes play a key role in these dynamics by shaping microbial response to the environment and among themselves. Nonetheless, bridging cellular metabolism to microbial self-organization phenomena is both conceptually and technically problematic. I will introduce these theoretical challenges, but also a self-organization phenomenon that offers a privileged position to solve the problem and preliminary modeling results. The presentation concludes with future perspectives.