Study Finds Ancestry Bias in CRISPR Screens

In the quest to unravel cancer's complexities, CRISPR/Cas9 gene editing has emerged as a powerful tool, enabling researchers to systematically deactivate genes and identify those crucial for cancer cell survival—the so-called "cancer dependencies." However, a recent study by scientists at the Broad Institute of MIT and Harvard has unveiled a concerning bias that could undermine the accuracy of these CRISPR screening experiments.

The culprit? CRISPR guides—short RNA sequences that direct the Cas9 enzyme to specific genomic sites for gene editing. Researchers found that about 2% of these guides fail to hit their intended targets, disproportionately affecting cells from individuals of African ancestry. This discrepancy stems from the fact that CRISPR guides were designed using reference genomes predominantly derived from European populations, failing to account for global genetic diversity.

"These inaccuracies exist in places we might not recognize and in ways that we wouldn't have predicted," said Rameen Beroukhim, co-senior author of the study published in Nature Communications. "This work shows that it's really worthwhile to conduct a systematic assessment of all the tools and datasets that we're using so that we can fix these hidden biases before they become an issue."

The team's analysis of the Broad's Cancer Dependency Map (DepMap), the largest cancer dependency resource, revealed that while less than 1% of cell line-guide pairs were affected, these biases are significant and must be addressed. In response, the DepMap team removed all guide RNAs that didn't work, ensuring that the database no longer falsely reports no dependencies for affected genes.

Sean Misek, the study's first author, initially set out to explore how germline genetic variants influence tumor response to treatment. His findings revealed strong associations between ancestry and genetic dependencies, with most stemming from artifacts related to germline variants and mismatches between guide RNA sequences and target genetic sequences. 

"These sorts of experimental biases are probably everywhere in preclinical research," Misek said. "We hope that this paper is part of a larger conversation."

To aid researchers in mitigating this issue, the team developed Ancestry Garden, a website that leverages data from the Genome Aggregation Database (gnomAD) to help determine the effect of ancestry on a chosen guide.

While the impact of this bias was relatively modest in the DepMap, it may be much larger in experiments studying only one or a small number of cell lines. To address this, the study team and DepMap researchers emphasize the importance of increasing genetic diversity in large-scale cell line libraries.

"We encourage the community to send us cell lines from under-represented populations if they have them," said Francisca Vazquez, director of the DepMap at the Broad. "This is a very important issue to address."