McVicker, G. | Search bioRxiv

bioRxiv2019

Discovering functional sequences with RELICS, an analysis method for CRISPR regulatory screens

CRISPR screens are a powerful new technology for the identification of genome sequences that affect cellular phenotypes such as gene expression, survival, and proliferation. By tiling single-guide RNA (sgRNA) target sites across large genomic regions, CRISPR screens have the potential to systematically discovery novel functional sequences, however, a lack of purpose-built analysis tools limits the effectiveness of this approach. Here we describe RELICS, a Bayesian hierarchical model for the discovery of functional sequences from tiling CRISPR screens. RELICS considers the overlapping effects of multiple nearby functional sequences, accounts for the area of effect surrounding sgRNA target sites, models overdispersion in sgRNA counts, combines information across multiple pools, and estimates the number of functional sequences supported by the data. In simulations, RELICS outperforms existing methods and provides higher resolution predictions. We apply RELICS to published CRISPR interference and CRISPR activation screens and predict novel regulatory sequences, several of which we experimentally validate. In summary, RELICS is a powerful new analysis method for tiling CRISPR screens that enables the discovery of functional sequences with unprecedented resolution and accuracy.

bioinformatics↗

bioRxiv2019

BreakCA, a method to discover indels using ChIP-seq and ATAC-seq reads, finds recurrent indels in regulatory regions of neuroblastoma genomes

Most known cancer driver mutations are within protein coding regions of the genome, however, there are several important examples of oncogenic non-coding regulatory mutations. We developed a method to identify insertions and deletions (indels) in regulatory regions using aligned reads from chromatin immunoprecipitation followed by sequencing (ChIP-seq) or the assay for transposase-accessible chromatin (ATAC-seq). Our method, which we call BreakCA for Breaks in Chromatin Accessible regions, allows non-coding indels to be discovered in the absence of whole genome sequencing data, out-performs popular variant callers such as the GATK-HaplotypeCaller and VarScan2, and detects known oncogenic regulatory mutations in T-cell acute lymphoblastic leukemia cell lines. We apply BreakCA to identify indels in H3K27ac ChIP-seq peaks in 23 neuroblastoma cell lines and, after removing common germline variants, we identify 23 rare germline or somatic indels that occur in multiple neuroblastoma cell lines. Among them, 4 indels are candidate oncogenic drivers that are present in 4 or 5 cell lines, absent from the genome aggregation database of over 15,000 whole genome sequences, and within the promoters or first introns of known genes (PHF21A, ADAMTS19, GPR85 and RALGDS). In addition, we observe a rare 7bp germline deletion in two cell lines, which is associated with high expression of the histone demethylase KDM5B. Overexpression of KDM5B is prognostic for many cancers and further characterization of this indel as a potential oncogenic risk factor is therefore warranted.

bioinformatics↗