Journal: Scientific Reports
Article Title: Nanopore signal deviations from pseudouridine modifications in RNA are sequence-specific: quantification requires dedicated synthetic controls
doi: 10.1038/s41598-024-72994-9
Figure Lengend Snippet: Quantitative profiling of ψ sites in human mRNAs from different cell lines. ( A ) Predicted site-specific ψ occupancies on human mRNA transcripts sequenced from A549 (black, n = 4), HepG2 (orange red, n = 3), HEK293T (blue, n = 2), Hct116 (orange, n = 2), HeLa (green, n = 3), HeLa TRUB1 KD (red markers inside yellow boundaries, n = 1), NTERA (purple, n = 2), and SH-SY5Y (cyan, n = 3) cells using a GBC model (see Methods). Each marker shows the mean and standard deviation of the ψ occupancy predicted by the GBC models trained for a specific site on the mRNA of a gene in an independent sequencing library. The box and whiskers plots show the combined GBC predictions of all independent sequencing libraries per cell line. The number of reads extracted from each gene in the associated cell line for GBC ψ-prediction is annotated, and the order of the read sample size corresponds to the markers displaying the results for each sequencing library. *Average ψ occupancy results reported by recently published bisulfite sequencing methods ‘PRAISE’ (black square) and ‘BID-seq’ (pink triangle) for PSMB2 and MCM5 mRNA transcripts in HEK293T are displayed inside blue boundaries for comparison. Mean and standard deviation values for each replicate are provided in Supplemental Table . ( B ) Heatmap of the statistical significance, i.e., \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$-10\text{log}[p\text{-}value]$$\end{document} , of the difference in ψ occupancy between cell lines for a given gene where p < 0.01 corresponds to a p -value greater than 30. For each cell line, a p -value was determined through an independent two-sample t-test, with one sample set consisting of the cell line’s predicted ψ fractions (box and whiskers in panel a), and the other sample set comprising the predicted ψ fractions for all other cell lines. ( C ) Comparison of the standard deviations for the ψ TP rate when PSMB2 synthetic ψ-modified transcripts are predicted with either a U-to-C mismatch classification model (blue) or a GBC model (green) as a function of read coverage. Standard deviation of the TP rate for both classifiers is calculated by taking their re-corrected TP rate, where the re-correction fits are calculated from the TP and FP trends in Fig. A–E (see Methods for details), at each read coverage increment on the x-axis (ranging from n = 7 to n = 200 in increments of 2) for all 25 model iterations (both GBC and U-to-C classifier) used in the analysis shown in Fig. A. The gold histogram in the background displays the total differential mRNA read coverage captured in one of the libraries sequenced from HeLa cells (median coverage = 10 reads). Square markers indicate actual standard deviations for ψ occupancy predicted for PSMB2 in three biological replicates of HeLa DRS.
Article Snippet: HeLa cervical adenocarcinoma cells (ATCC CCL-2), A549 pulmonary carcinoma cells (ATCC CCL-185), NTERA testicular malignant pluripotent embryonal carcinoma cells (ATCC CRL-1973), and HepG2 liver hepatocellular carcinoma cells (ATCC HB-8065) were cultured in Dulbecco’s modified Eagle’s medium (Gibco, 10566024), supplemented with 10% Fetal Bovine Serum (FB12999102, Fisher Scientific) and 1% Penicillin–Streptomycin (Lonza,17602E).
Techniques: Marker, Standard Deviation, Sequencing, Methylation Sequencing, Comparison, Modification
ATCC is a verified supplier 
ATCC manufactures this product 
![Quantitative profiling of ψ sites in human mRNAs from different cell lines. ( A ) Predicted site-specific ψ occupancies on human mRNA transcripts sequenced from A549 (black, n = 4), HepG2 (orange red, n = 3), HEK293T (blue, n = 2), Hct116 (orange, n = 2), HeLa (green, n = 3), HeLa TRUB1 KD (red markers inside yellow boundaries, n = 1), <t>NTERA</t> (purple, n = 2), and SH-SY5Y (cyan, n = 3) cells using a GBC model (see Methods). Each marker shows the mean and standard deviation of the ψ occupancy predicted by the GBC models trained for a specific site on the mRNA of a gene in an independent sequencing library. The box and whiskers plots show the combined GBC predictions of all independent sequencing libraries per cell line. The number of reads extracted from each gene in the associated cell line for GBC ψ-prediction is annotated, and the order of the read sample size corresponds to the markers displaying the results for each sequencing library. *Average ψ occupancy results reported by recently published bisulfite sequencing methods ‘PRAISE’ (black square) and ‘BID-seq’ (pink triangle) for PSMB2 and MCM5 mRNA transcripts in HEK293T are displayed inside blue boundaries for comparison. Mean and standard deviation values for each replicate are provided in Supplemental Table . ( B ) Heatmap of the statistical significance, i.e., \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$-10\text{log}[p\text{-}value]$$\end{document} , of the difference in ψ occupancy between cell lines for a given gene where p < 0.01 corresponds to a p -value greater than 30. For each cell line, a p -value was determined through an independent two-sample t-test, with one sample set consisting of the cell line’s predicted ψ fractions (box and whiskers in panel a), and the other sample set comprising the predicted ψ fractions for all other cell lines. ( C ) Comparison of the standard deviations for the ψ TP rate when PSMB2 synthetic ψ-modified transcripts are predicted with either a U-to-C mismatch classification model (blue) or a GBC model (green) as a function of read coverage. Standard deviation of the TP rate for both classifiers is calculated by taking their re-corrected TP rate, where the re-correction fits are calculated from the TP and FP trends in Fig. A–E (see Methods for details), at each read coverage increment on the x-axis (ranging from n = 7 to n = 200 in increments of 2) for all 25 model iterations (both GBC and U-to-C classifier) used in the analysis shown in Fig. A. The gold histogram in the background displays the total differential mRNA read coverage captured in one of the libraries sequenced from HeLa cells (median coverage = 10 reads). Square markers indicate actual standard deviations for ψ occupancy predicted for PSMB2 in three biological replicates of HeLa DRS.](https://pub-med-central-images-cdn.bioz.com/pub_med_central_ids_ending_with_8862/pmc11438862/pmc11438862__41598_2024_72994_Fig4_HTML.jpg)