PON-Sol

PON-Sol predicts the effects of amino acid substitution on protein solubility. It is a machine learning-based method and utilizes amino acid features and evolutionary information. Please find the dataset for training PON-Sol here.

The method is described in more detail in PON-Sol: prediction of effects of amino acid substitutions on protein solubility.

Variations predict Protein Predict About Disclaimer

PON-Sol

PON-Sol is a machine-learning based tool that predict the effect of variation on protein solubility. It is trained using random forest algorithm and utilizes amino acid features and evolutionary information.


Training dataset used in PON-Sol

The training dataset containing 399 varaiants in 87 proteins is collected from literatures. It can be downloaded from here.


Submit queries to PON-Sol

Variations

Multiple variations in a single protein or in multiple proteins can be submitted as a single query. A file containing one variation and matched protein gi number in each line should be uploaded.

Format Example:
L723P   4502167
V717F   4502167

Protein

You can also submit a protein gi to find all possible variations' effect on protein solubility.

Email
Email field is obligatory. The predictions will be sent to you in the provided email address.


PON-Sol output

An email with the result file(s) as an attachment will be sent to the email address provided during job submission. The result file contains the following contents:
Prediction
1. Amino acid substitution(s)
2. Protein gi(s)
3. Solubility prediction result: increase/decrease/no-change
4. Score of PON-sol(probability): range from 0 to 1

If the user submit a protein gi, the result will have two PDF files:
1. Top 25 variations that induce protein solubility to increase and decrease seperately.
2. All possible variations's effects to solublity in every position of the protein.

How to cite PON-Sol?

Yang Yang, Abhishek Niroula, Bairong Shen, and Mauno Vihinen
PON-Sol: prediction of effects of amino acid substitutions on protein solubility
Bioinformatics 2016 32(13): 2032-2034.


If you have any queries, please feel free to contact us.