“…Shambhala‐1 allows one‐by‐one adding of new harmonized expression samples to a common pre‐calculated pool, and no recalculation is needed for the whole set of samples. This feature can dramatically reduce calculation time and costs, which is especially sensitive considering next‐level gene expression metrics such as molecular pathway activation levels (Aliper et al., 2017; Borisov et al., 2017; Borisov, Sorokin, Garazha, & Buzdin, 2020; Buzdin, Prassolov, Zhavoronkov, & Borisov, 2017; Buzdin et al., 2014, 2018), individual drug sensitivity estimates (Poddubskaya et al., 2019; Tkachev, Sorokin, Garazha et al., 2020; Zolotovskaia et al., 2019), machine learning models (Borisov & Buzdin, 2019; Borisov et al., 2018, 2021; Tkachev et al., 2019; Tkachev, Sorokin, Borisov, et al., 2020), and more. Technically, the Shambhala‐1 method utilizes a piecewise linear normalization method (XPN) for building a universal gene expression matrix (Shabalin et al., 2008).…”