Fig. 3

Experimental validation of the CUT&RUN suspect lists. A Principal component analysis graphs (top row) and Scree plots (bottom row) of human and mouse negative control samples, showing a decrease in sample similarity after filtering with C&R suspect lists. The percentage of variance explained by each principal component decreases, and the samples become more randomly distributed on the graphs, indicating more variance in line with an expected random distribution of reads from a non-specific antibody in C&R. The Scree plots show a decrease in the Eigenvalue and contribution of principal component 1 to the total observed variance after filtering. B Principal component analysis (top row) and Scree plots (bottom row) on human and mouse negative controls performed to validate the C&R suspect lists, showing increase variance after filtering. C Comparison of C&R suspect lists with suspect lists built with the same method based on experiments performed in HEK293T human cells and embryonic tissues from JAX Swiss mice. The suspect lists show high concordance with over 80% of each C&R suspect list being identified by the cell or strain specific suspect lists. D Examples of genome coverage of representative samples under unique regions of the C&R and cell/strain specific suspect lists. C&R, CUT&RUN; PCA, principal component analysis