Explainable Generative Deep Learning for plant transcription factor binding sites discovery



Enter a FASTA sequence  
Paste an example




Minimum sequence length required: 160 bases








    

To generate AlphaFold2 PDB structure file for your amino acids sequence, please Click Here
Once generated, the PDB structure file can be uploaded above alongwith query DNA sequence to run PTF-Vāc



PTF-Vāc


According to the ancient and fundamental philosophy of India, precisely Upanishads and Trika, throb caused vibrations, vibration casued sound, the marker of creation, information. When this vibration is absorbed through knowledge, the utterences made thereafter is called Vāc.

PTF-Vāc has absorbed the vibrations of various exponents of DNA and transcription factors which allow their association and help recognize the binding sites across the DNA sequence, without needing any predefined transcription factor or species specific model or motif or matrix information.

Upto this time, finding transcription factors binding sites was deeply coupled with the process of motif discovery which essentially needed experimental binding data to draw the motif matrix, which could be used in turn to scan the DNA for binding sites. It meant that for any transcription factors and species for which such binding experiments were not done, one would not be able to know precisely the binding site for that TF in that species. In such scenario, information and matrices available for other species would be used, which is fundamentally wrong practise as in plants the TF binding sites vary a lot, and so do the TF structure itself.

PTF- Vāc has learned the structure and sequence co-variability through PTFSpot philosophy and applied a transformer Encoder:Decoder system to firmly place a system which can accurately tell the binding site for any TF in any given sequence, ab -initio, totally free of any prior species and transcription factor specific information and knowledges like matrices. It has decoupled the problem of binding site discovery from motif finding. The user can now place even a single sequence and can accuately identify the binding site for the selected TF, whether seen or known before or not for any plant genome, old or novel!


CITATION:


PTF-Vac: Ab-initio discovery of plant transcription factors binding sites using explainable and generative deep co-learning encoders-decoders
Sagar Gupta, Jyoti, Umesh Bhati, Veerbhan Kesarwani, Akanksha Sharma, Ravi Shankar* ;bioRxiv, 2024  
Read research article here
 

Credits: Image Source- Adobe Firefly