Download HumanBase predicted tissue-specific interaction networks
All networks are available as gzipped, tab-delimited text files. The file format is as follows:
Top Edges: the network is filtered to only include edges with evidence supporting a tissue-specific functional interaction
[entrez gene id 1][entrez gene id 2][posterior prob.]
Full Network: the network is fully connected
[entrez gene id 1][entrez gene id 2][posterior prob., with known edges set to 1][posterior prob.]
Gold Standard: each text file (e.g. kidney.dat) contains edges with values 1 through 4 corresponding to C1 - C4 described in our paper. C1 edges are used as positives and C2-C4 are used as negatives to train the tissue-specific models.
Note: These downloadable networks assume a prior probability of 0.1 to match the displayed networksBulk Downloads
Download gzip archives of all networks and standards. Note: These are very large files. Consider using the Python script below to download individual networks in smaller increments if your connection is unstable.
Bulk Download Script
Download tissue-specific networks using our Python script. This is recommended for downloading all networks incrementally or when dealing with unstable connections. Requirements: Python 3.6+ (no additional packages needed).
Individual Network Downloads
| Tissue | Top edges | Full network | Gold Standard |
|---|
Access pre-computed variant effect predictions and model weights from HumanBase's suite of genomic interpretation tools
ExPecto Resources
ExPecto predicts tissue-specific expression effects of genetic variants from sequence, providing quantitative predictions for how mutations affect gene expression across 218 human tissues and cell types.
1000 Genomes SNP predictions
File: combined_snps.0.3.zip
Size: 1.9 GB
Format: ZIP archive
1000 Genomes variants that passed a minimum predicted effect threshold (>0.3 log fold-change in any tissue).
Download 1000 Genomes SNP PredictionsAll 1kb Mutations
File: all1kbmutations.tar
Size: 128 GB
Format: TAR archive
Full prediction of all 140 million mutations within 1kb to TSS across the human genome.
Download All 1kb MutationsEvolutionary Constraint Analysis
File: evocon.zip
Size: 34.4 MB
Format: ZIP archive
Variation potential directionality scores and inferred evolutionary constraint probabilities for genes across tissues. Negative scores indicate constraint toward higher expression.
Download Evolutionary Constraint DataExPectoSC Resources
ExPectoSC extends ExPecto's capabilities to predict cell-type-specific chromatin effects and expression changes at single-cell resolution.
ClinVar Scaled Predictions
File: clinvar_1000G_final_nc_all_info.csv
Size: 114.4 MB
Format: CSV
ExPectoSC predictions for ClinVar variants within +/- 20kb of representative TSS, with z-score normalization using 1000 Genomes variants.
Download ClinVar PredictionssLDSC Annotations
File: CLEVER_preds_sLDSC_annot.tgz
Size: 2.0 GB
Format: TGZ (compressed TAR archive)
Stratified Linkage Disequilibrium Score regression (sLDSC) annotations derived from ExPectoSC predictions.
Download sLDSC AnnotationsSei Resources
Sei is a deep learning framework for systematically predicting sequence regulatory activities and understanding human genetics data through comprehensive chromatin profiling predictions.
Sei Model Resources
File: resources.tar.gz
Size: 3.4 GB
Format: TAR GZ archive
Pre-computed resources and reference data for Sei chromatin profiling predictions and regulatory activity analysis.
Download Sei ResourcesSeqweaver Resources
Seqweaver predicts how genetic variants affect post-transcriptional RNA-binding protein (RBP) interactions using deep learning models trained on RBP-RNA interaction data from CLIP-seq experiments.
1000 Genome Project RBP LDScores
File: Seqweaver_RBP_ldscores.tar.gz
Size: 33.5 GB
Format: TAR GZ archive
RBP LDScores for GWAS analysis using the 1000 Genome Project data.
Download RBP LDScoresGnomAD RBP Target Site Dysregulation Scores
File: Seqweaver_RBP_gnomAD.tar.gz
Size: 37.6 GB
Format: TAR GZ archive
GnomAD (v2.1) Seqweaver RBP target site dysregulation scores.
Download GnomAD RBP ScoresSeqweaver RBP Model
File: Seqweaver-v0.1.tar.gz
Size: 6.6 GB
Format: TAR GZ archive
Code for making variant effect predictions using Seqweaver RBP models. Also used as the RNA model in the ASD Browser for predicting molecular-level effects of SSC ASD proband mutations.
Download Seqweaver ModelDeepSEA/Beluga Resources
DeepSEA Beluga is a deep learning framework that predicts genomic variant effects on chromatin features using convolutional neural networks trained on large-scale chromatin profiling data.
Model Weights
File: deepsea.beluga.pth
Size: 570.4 MB
Format: PyTorch model file (.pth)
Pre-trained weights for the DeepSEA "Beluga" model, which predicts 2002 chromatin features including transcription factor binding, DNase I sensitivities, and histone marks.
Download DeepSEA Beluga ModelASD Browser Resources
Access the code for making variant effect predictions from the DNA and RNA models used in the ASD Browser analysis.
DNA and RNA Models
File: code_asd_dnarna_v3.tar.gz
Size: 2.19 GB
Format: TAR GZ archive
Code for making variant effect predictions from the DNA and RNA models used in the ASD Browser analysis described in “Whole-genome deep learning analysis reveals causal role of noncoding mutations in autism”.
Download DNA and RNA Model Code
Citation
When using these resources, please cite the relevant publications:
- ExPectoSC: Sokolova, K., Theesfeld, C. L., Wong, A. K., Zhang, Z., Dolinski, K., & Troyanskaya, O. G. (2023).
Atlas of primary cell-type specific sequence models of gene expression and variant effects. Cell Reports Methods. - Sei: Chen, K. M., Wong, A. K., Troyanskaya, O. G., & Zhou, J. (2022).
A sequence-based global map of regulatory activity for deciphering human genetics. Nature Genetics. - Seqweaver: Park, C. Y., Zhou, J., Wong, A. K., Chen, K. M., Theesfeld, C. L., Darnell, R. B., & Troyanskaya, O. G. (2021).
Genome-wide landscape of RNA-binding protein target site dysregulation reveals a major impact on psychiatric disorder risk. Nature Genetics. - ASD Browser: Zhou, J., Park, C. Y., Theesfeld, C. L., Wong, A. K., Yuan, Y., Scheckel, C., Fak, J. J., Funk, J., Yao, K., Tajima, Y., Packer, A., Darnell, R. B., & Troyanskaya, O. G. (2019).
Whole-genome deep-learning analysis identifies contribution of noncoding mutations to autism risk. Nature Genetics. - ExPecto: Zhou, J., Theesfeld, C. L., Yao, K., Chen, K. M., Wong, A. K., & Troyanskaya, O. G. (2018).
Deep learning sequence-based ab initio prediction of variant effects on expression and disease risk. Nature Genetics. - DeepSEA (Beluga): Zhou, J., & Troyanskaya, O. G. (2015).
Predicting the Effects of Noncoding Variants with Deep learning-based Sequence Model. Nature Methods.