Abstract Enzyme-mediated chemical modifications to mRNA are important for fine-tuning gene expression, but they are challenging to quantify due to low copy number and limited tools for accurate detection. Existing studies have typically focused on the identification and impact of adenine modifications on mRNA (m 6 A and inosine) due to the availability of analytical methods. The pseudouridine (Ψ) mRNA modification is also highly abundant but difficult to detect and quantify because there is no available antibody, it is mass silent, and maintains canonical basepairing with adenine. Nanopores may be used to directly identify Ψ sites in RNAs using a systematically miscalled base, however, this approach is not quantitative and highly sequence dependent. In this work, we apply supervised machine learning models that are trained on sequence-specific, synthetic controls to endogenous transcriptome data and achieve the first quantitative Ψ occupancy measurement in human mRNAs. Our supervised machine learning models reveal that for every site studied, different signal parameters are required to maximize Ψ classification accuracy. We show that applying our model is critical for quantification, especially in low-abundance mRNAs. Our engine can be used to profile Ψ-occupancy across cell types and cell states, thus providing critical insights about physiological relevance of Ψ modification to mRNAs.