Fig. 1

Diverse MAVEs and the datasets they produce. a DMS assays using either affinity purification or selective growth. (i) The DMS assay of Olson et al. [8] used a library of variant GB1 proteins physically linked to their coding mRNAs. Functional GB1 proteins were then enriched using IgG beads. (ii) The DMS studies of Seuma et al. [9] and Bolognesi et al. [10] used selective growth in genetically modified Saccharomyces cerevisiae to assay the functionality of variant A β and TDP-43 proteins, respectively. In all three experiments, deep sequencing was used to determine an enrichment ratio for each protein variant. b The resulting DMS datasets consist of variant protein sequences and their corresponding log enrichment values. c The MPSA of Wong et al. [11]. A library of 3-exon minigenes was constructed from exons 16, 17, and 18 of the human BRCA2 gene, with each minigene having a variant 5 ^′ss at exon 17 and a random 20 nt barcode in the 3 ^′ UTR. This library was transfected into HeLa cells, and deep sequencing of RNA barcodes was used to quantify mRNA isoform abundance. d The resulting MPSA dataset comprises variant 5 ^′ss with (noisy) PSI values. e The sort-seq MPRA of Kinney et al. [12]. A plasmid library was generated in which randomly mutagenized versions of the Escherichia coli lac promoter drove the expression of GFP. Cells carrying these plasmids were sorted using FACS, and the variant promoters in each bin of sorted cells, as well as the initial library, were sequenced. f The resulting dataset comprises a list of variant promoter sequences, as well as a matrix of counts for each variant in each FACS bin. MAVE: multiplex assay of variant effect; DMS: deep mutational scanning; GB1: protein G domain B1; IgG: immunoglobulin G; A β: amyloid beta; TDP-43: TAR DNA-binding protein 43; MPSA: massively parallel splicing assay; BRCA2: breast cancer 2; 5 ^′ss: 5 ^′ splice site(s); UTR: untranslated region; PSI: percent spliced in; GFP: green fluorescent protein; FACS: fluorescence-activated cell sorting