ePOSSUM2 Predicting the impact of DNA variants on transcription factor binding. ePOSSUM2
1. Enter variant
Variant (VCF like)

17:19437135G>A or 17:19437134GG>G
Positions must refer to genome build 37.
OR
Wild-type sequence

GGCCCTCACACTCTCCAACCTCATCTCCCTGGT
Variant sequence

TCACACTCTCCAACCTCATCTCCCTGGTGAG
OR
run a batch query

3. Search for binding sites
2. Select transcription factors


Documentation

Simply enter the DNA variant you would like to analyse into the variant field, select one or multiple transcription factors and click on submit. If you do not know the location but have a wild-type and a variant sequence, you can still enter them. In this case, ePOSSUM2 cannot display any known transcription binding sites.

There is detailed documentation of ePOSSUM2's output.
We also provide a disease-linked real-life example.

ePOSSUM2 currently only works with human transcription factors.
About ePOSSUM2
ePOSSUM2 supports more than 1,000 different binding models for 247 human transcription factors. The models were pooled from various publicly accessible data sources, such as CIS-BP, Hocomoco, hPDI, JASPAR and Jolma (2013). 103 transcription factor flexible models (TFFMs) are included. For each binding model, ePOSSUM2 constructs a Bayes classifier and reports the likelihood that a variant leads to the gain or loss of a transcription factor binding site. Because of the poor performance of several binding models in detecting experimentally verified sites, a negative or positive predictive value is reported.
Download the underlying data.

ePOSSUM2 has been developed in 2018 at Berlin Institute of Health (BIH) and Charité – Universitätsmedizin Berlin by Robin Steinhaus, Peter N. Robinson and Dominik Seelow. ePOSSUM2 is an update of the ePOSSUM software. ePOSSUM2 is free to use.

Imprint / privacy