IE-Tx: analysis code and result tables for a sequence-feature analysis of TBK1-responsive genes in the planarian Schmidtea mediterranea

Analysis code, feature tables, label sets and result tables for a study of thesequence features that distinguish genes down-regulated upon TBK1 knockdown inthe planarian Schmidtea mediterranea. The study asks whether a sequence-based classifier trained on this label setreads biology or the technical correlates of RNA-seq detectability. Nineinterpretable genomic features were extracted and evaluated underhomology-cluster-aware cross-validation, cluster-level bootstrap and permutationtesting. The cluster-wise AUC is 0.725 (permutation p = 0.0010); rebuilding thelabels on the two-sided criterion used for the external perturbations gives0.673. The readable signal is nucleotide composition and RNA secondarystructure, not immune-specific motifs. Contents: 27 main analysis scripts, 8 supplementary scripts, 4 figure scripts,51 result tables, feature tables, a gene-to-homology-cluster map (919 clusters,4373 genes), 7 candidate constructs and validation tables. Raw data are public and must be downloaded separately: genome and annotationfrom Zenodo record 13798866, TBK1-knockdown RNA-seq from Zenodo record 17072252. Code: MIT. Data tables: CC BY 4.0.