Extended Data Fig. 7: Comparison between RNA Immunoprecipitation Sequencing (RIP-seq) and PRIM-seq data.
From: Genome-wide mapping of RNA-protein associations through sequencing
(a) Confusion matrix of all human RNA genes categorized according to their identification as SMC1A targets by PRIM-seq (rows) and RIP-seq (columns). (b) Chi-square test results derived from the confusion matrix in (a). (c–f) ROC curves evaluating PRIM-seq data against RIP-seq results for SMC1A, SMC3, RAD21, and HDAC2. RIP-seq-identified targets serve as the reference sets, and varying thresholds on PRIM-seq read counts are applied to define PRIM-seq-identified targets. (g–j) Expanded views of the ROC curves in (c–f), highlighting lower values on the x-axis. (k) Confusion matrix based solely on PRIM-seq reads within the HuRPA network, categorizing reads according to whether the protein-end maps to SMC1A (rows) and whether the RNA-end corresponds to SMC1A RIP-seq–identified target RNAs (columns). (l) Chi-square test results from the confusion matrix shown in (k). (m) Chi-square test results from a restricted version of the confusion matrix in (k), using only PRIM-seq reads involving SMC1A, SMC3, RAD21, and HDAC2 and their respective RIP-seq–identified target RNAs.
