Prediction Tool
Prediction model to accurately forecast the observed rate constant of DNA sequences. By inputting specific parameters such as the DNA sequence, concentrations of NaCl and KCl, cofactor element type and concentration, desired pH, and temperature values, users can obtain a canalysis of the sequence's behavior under designated buffer conditions.
Frequently Asked Questions
The activity prediction process leverages a curated database of RNA-cleaving catalytic cores, utilizes novel autoencoder-derived and 2-mer descriptors, includes buffer and cofactor representation, and employs the LightGBM model to estimate rate constant values.
The prediction accuracy is high with a Q2 value of 0.93 and RMSE of 0.52, showing the model's capability to estimate observed rate constant with an error of 1/3 order of magnitude across a wide range from 0.0001 to 1.7 min-1.
Please email us a concise description of your experimental results, including the catalytic center sequence and experimental conditions. Our team will review and incorporate your submission into our platform and model, allowing others to benefit from and build upon your research.