Single-cell sequencing techniques allow the study of the subclonal architecture of tumours and reveal the co-occurrence of (driver) mutations as well as their order of acquisition over molecular pseudo-time. The development of multi-omics techniques such as genome-and-transcriptome sequencing (G&T-seq) will enable us to study the diversity of cancer cell states that arises within a tumour, at its most fundamental level, the cell. Recently we developed Gtag&T-seq, an improved G&T-seq method that applies direct genome tagmentation (Gtag), thereby avoiding upfront whole-genome amplification (WGA) and its associated cost and biases. We applied both methodologies to a human melanoma PDX melanoma model before treatment with BRAF and MEK inhibition, and at minimal residual disease. Three genomic subclones were observed at both timepoints, however their relative abundancy at cellular level changes upon drug exposure. Using the RNA of the same single cells, we showed that transcriptome-based DNA copy number inference has limited resolution and accuracy, underlining the importance of affordable multi-omic approaches. In addition, we discerned 10 focal amplifications on chromosomes 13 and 22 of which only genes of the chr22q11.21 focal amplification showed clear gene dosage effects. Using the Gtag&T data, we were able to determine breakpoints at near base-pair resolution, which is not possible with conventional G&T-seq. Finally, we mapped known melanoma resistant cell states on top of the phylogenetic cell lineage tree revealing that some cell states do not occur in all of the subclones. In conclusion, Gtag&T-seq is a novel, low-cost, and accurate single-cell multi-omics method that enables the exploration of somatic genetic alterations and their functional consequences in single cells at scale.