DNA recognition preference of Transcription factors through co-evolutionary modeling
(Coevolved-TF-DNA)

Scan sequences for TF binding

A brief overview of the mfDCA process is explained in "Direct-coupling analysis of residue coevolution captures native contacts across many protein families" by Faruck Morcos et al. and in "Global pairwise RNA interaction landscapes reveal core features of protein recognition" by Qin Zhou et al. This implementation focuses on predicting transcription factor binding affinities to user-submitted DNA sequences.

How to use this tool:
1. Select the Transcription Factor of interest from the drop-down list.
2. Select which input method to use with the bubbles next to each option, then paste or upload (.fasta) your sequence accordingly.
3. Input your desired number of hits, leaving the value as 0 for none to be displayed.
4. Click the submit button


Basic Options
Select Transcription Factor: *
(Human genes are in all caps)
(Mouse genes are capitalized)


 Select input type: *
(Sequences must be between 20nt and 1,000,000nt)
Display options: *
(Values must be between 0 - 100 or 0.1% - 5%)
(If percent value exceeds 100 hits, only 100 hits will be shown)

Display the or percent most likely binding results.
Display the or percent most unlikely binding results.

(Expect a delay of 20-30 seconds)


Advanced Options

Display only the table(s) for Hamiltonian scores

Display top scores as graphs with nearby sequences, combining nearby peaks into single plots.
Graph and of a result.

Display a single plot of all Hamiltonian scores for the full sequence (recommended only for <4000nt).

Final Hamiltonian scores are displayed along with their 20mer sequence, position in the original sequence, and the name of their sequence of origin. If the fasta file contains only a single sequence, then the origin can be safely omitted.

Directly output & download the full list of hamiltonian scores to file, rather than displaying them.

Genomic binding sites