Data repository associated with 'A Functional Map of the Human Intrinsically Disordered Proteome'

ES_MAP.zip

• a hierarchically clustered map of the human IDR-ome
• .cdt .gtr files - outputs of Cluster3.0 software 
• .txt file equivalent of the .cdt
• can be visualized using JavaTreeView (see Tutorial_ES.pdf)

TUTORIAL.zip, information on:

• visualization and analysis of the human IDR-ome map
• search for proteins of interest and exploratory analyses of clusters
• automatic export and analysis of exported clusters (code available at https://github.com/IPritisanac/ES_PW)

IDROME_SEQUENCES.zip

• human proteome fasta file
• IDRome fasta file
• SPOT-Disorder v1.0 disorder boundaries
  • 13 044 unique protein sequences with at least one IDR (>=30 amino acids)
  • 21 252 total unique human IDRs

IDR_ALN.zip

• alignments of IDR sequences across ENSEMBL orthologs
• 19 459 IDR alignments
• UniProt ID and IDR boundaries for the human sequence are indicated in the name of the file

FAIDR_TSTATS.zip

• hierarchical clustering of FAIDR t-statistics for 148 GO terms
  
  • .cdt, .gtr files from Cluster3.0
  • can be visualized using JavaTreeView
  • reveals the most predictive molecular features for the top performing 148 models

CLUSTERS_EXPLORE.zip

• clusters obtained through exploratory analysis of the map provided in ES_MAP.zip
• 93 exported clusters in .cdt file format

CLUSTERS_AUTO.zip

• clusters extracted from the hierarchically clustered IDR-ome map at a range of distance thresholds (0.4 - 0.8) in .cdt file format
• distance refers to the uncentered correlation distance between vectors of Z-scores representing human IDRs
• clusters extracted at different distance thresholds are split into separate archives
• AUTO_GO_FEATS.xlsx - summary of GO-term overrepresentation and feature enrichment analyses; each distance threshold is in a separate sheet

FAIDR_HIGH_AUC_PPV_GO.zip

• target files with annotations of 148 GO terms for which good quality FAIDR models could be obtained (AUC >= 0.7, PPV >= 0.4)
• file format: three columns; 1st: IDR ID (includes IDR boundaries); 2nd: protein UniProt ID; 3rd: annotation of the protein to a GO term (1 if known to be associated with the GO term, 0 if not)

DATASETS.zip

• supplementary dataset accompanying manuscript (PNAS 2026)
  
  

FS_UP000005640_9606.zip

• Per-IDR-sequence mean Z-scores for all features. These feature Z-scores are computed from 'raw' feature values and normalized by the global mean and standard deviation of the features across all IDRs in the human proteome.