Fig 1.
U2OS cells prepared for this study were stained using the Cell Painting assay protocol [34], with six stains imaged across five channels, revealing eight cellular components/structures. Scale bar 25 μm.
Fig 2.
Workflow for generating morphological profiles using Cell Painting.
U2OS cells are treated with shRNAs and transferred to 384-well plates in which they are stained and then imaged. The images are analyzed and ~1400 features are extracted from each cell. These data are then transformed to generate multivariate profiles.
Fig 3.
Different RNAi reagents targeting the same gene rarely produce similar profiles, whereas RNAi reagents sharing a seed sequence do.
shRNAs targeting the same gene have very low correlation (B), whereas those containing the same seed sequence have a much higher correlation (C). Using the 95th percentile of a null distribution (A) as a threshold to define significant correlations, only 10% of correlations in B are significant, compared to 73% in C. This indicates that the phenotypes induced by RNAi knockdown are dominated by seed effects. Correlations are computed between profiles of sequences, obtained by median-averaging profiles of replicate wells. The percentage of correlations above the defined threshold is indicated for each group; dotted line indicates 95th percentile of the null distribution (A). The difference between means of B and C is highly significant (P-value < 10−5; two-sided Student's t-test).
Fig 4.
Image-based profiles of RNAi sequences are highly reproducible.
Using the 95th percentile of the null distribution (A) as a threshold to define significant correlations, 92% of replicate correlations in B are seen to be significant. Correlations are computed between profiles of individual wells. The percentage of correlations above the defined threshold is indicated; dotted line indicates 95th percentile of the null distribution (A). The difference between means of A and B is highly significant (P-value < 10−15; two-sided Student's t-test).
Fig 5.
Visual example of seed effects dominating morphological profiles.
Only nuclear shape features were used in this example in order to yield visually interpretable phenotypes. Two of the shRNA sequences targeting CDCA7L and BECN1 share a seed sequence (sequences 299864 and 17937 have a common seed GAATGA at nucleotides 12–17). The Spearman correlation between the morphological profiles of these shRNAs is high (seed correlation = 0.44, red bar). For each of these two shRNAs, a same-gene shRNA with a dissimilar phenotype is also shown (same-gene shRNA correlation = -0.31 and -0.11, green bars). All four shRNAs have high replicate correlation and are dissimilar from untreated cells. We specifically chose an example where the seed correlation is high and the same-gene shRNA correlations are low; however this phenomenon is seen globally (Fig 3). Images have been zoomed in, showing only 2% of the imaged region.