Cell culture media composition and culture conditions play a crucial role in product yield, quality and cost of production. Culture media optimization is the technique of improving media composition and culture conditions to achieve desired product outcomes. To achieve this, there have been many algorithmic methods proposed and used for culture media optimization in the literature. To help readers evaluate and decide on a method that best suits their specific application, we carried out a systematic review of the different methods from an algorithmic perspective that classifies, explains and compares the available methods. We also examine the trends and new developments in the area. This review provides recommendations to researchers regarding the suitable media optimization algorithm for their applications and we hope to also promote the development of new cell culture media optimization methods that are better suited to existing and upcoming challenges in this biotechnology field, which will be essential for more efficient production of various cell culture products.

In recent years, supervised machine learning models trained on videos of animals with pose estimation data and behavior labels have been used for automated behavior classification. Applications include, for example, automated detection of neurological diseases in animal models. However, there are two problems with these supervised learning models. First, such models require a large amount of labeled data but the labeling of behaviors frame by frame is a laborious manual process that is not easily scalable. Second, such methods rely on handcrafted features obtained from pose estimation data that are usually designed empirically. In this paper, we propose to overcome these two problems using contrastive learning for self-supervised feature engineering on pose estimation data. Our approach allows the use of unlabeled videos to learn feature representations and reduce the need for handcrafting of higher-level features from pose positions. We show that this approach to feature representation can achieve better classification performance compared to handcrafted features alone, and that the performance improvement is due to contrastive learning on unlabeled data rather than the neural network architecture.

No abstract