genomics - functional assays I technology

why not just measure naturally occurring molecular species?

many important natural genetic variants are rare - it may be hard to find individuals w them. 60% of variants are unique among 1 million people

making perturbations is a powerful way to test hypotheses

depp learning: we often need bigger, more informative training sets to model cellular systems

how do we leverage the power of cheap sequencing/synthesis to perform functional assays at scale?

synthesize a library of variants/cCREs, gRNAs

turn your measurement of function into a sequencing assay

barcoding system

cis regulatory element (cre) or whats being tested for functionality

minimal promoter

reporter gene

unique dna barcode

sort-seq

reporter gene = gfp

flow cytometry - sort into bins of fluorescence level, sequence barcodes to ID calls in each bin

landing pads

single copy genome integrated construct for recombination-mediated integration of dna libraries

has recombinase sites so it can be integrated into a defined locus w crispr

requires prior cell line engineering

library recombination efficiency into landing pad can be low

one library member per cell at a defined location

can be used to test same cre in diff chromosomal locations

plamids

easy to clone libraries

requires transfectable cell type

multiple library members per cell

lentivirus/adeno-assoicated virus

genome integrated

genomic positioin can affect expression of construct

about one library member per cell when applied at low multiplicity of infection (moi)

construct size may be limited by what can be packaged in virus

key tricks for functional assays

you need an assay to directly test the function of non-coding sequences & variants at scale:

turn your functional assay into a sequencing assay

leverage cheap dna synthesis to build libraries of designed sequences

use sequence barcodes to id library elements in readout

massively parallel reporter assay (mpra)

pooled library of lots of distinct barcoded reporter genes

libraries can be made by custom array oligo synthesis

can be used to test cres, enhancers, splicing, utrs, etc

self-transcribing active regulatory region sequencing (starr-seq)

sheared genomic dna creates genomic fragments that can be cloned downstream of reporter gene, it itself acts as an enhancer & barcode

mpra limitations

lack of native genomic context

multiple dna copy number

mainly done in transfectable cell lines which may not recapitulate enhancer biology

potential assay toxicity: plasmids, high reporter gene levels

compatibility w reporter/minimal promoter: minimal promoter used may lead to false positives or negatives

VUS

variants of unknown significance

solving their effects very impactful for cancer research

deep mutational scans

create a library of all possible protein aa variants

pick an assayable function

pick a cell type

integrate one variant per cell

read out assay by sequencing dna of variants

variant abundance by massively parallel sequencing (vamp-seq)

create variants by site-saturation mutagenesis

variants fused to gfp, sequence barcodes cloned in

integrate library at single locus in cells

sort seq

what you need for pooled crispr screen

a library of designed guide rnas

a common readout of function (maybe reporter)

method to get cas9 and 1 grna per cell

a way to link grna identity w phenotypic readout of the cell

perturb-seq

combine pooled crispr library w single cell rna-seq