Yes it is possible

Sequence-based Motif :

Serine/Threonine kinases (e.g., PKA, AKT) often target sequences like R-X-S/T-X.

Tyrosine kinases (e.g., EGFR, Src) recognize sequences like Y-X-X-X-Y.

Tools like MotifX, ScanSite, and PhosphoSitePlus can help identify these motifs in a given protein sequence. These tools scan the protein's sequence for possible phosphorylation sites that match the kinase's recognition motif.

Data-base and model :

-PhosphoSitePlus: Contains a large collection of experimentally validated phosphorylation sites and kinase-substrate interactions.

-KINOMESCAN: A resource for identifying kinase-specific substrates.

NetPhos: A tool for predicting phosphorylation sites on serine, threonine, and tyrosine residues based on sequence.

-KinasePhos is one example of a prediction tool that uses machine learning to predict phosphorylation sites.

-PredKin and KINASEPRED are other tools that leverage sequence-based features and machine learning to predict kinase-substrate interactions.

Structural approches :

- alphaFold, PyMol These can be used to predict the 3D structures of the protein substrates and kinases, which can then be analyzed for potential binding interfaces.

-DynaMine: A tool for dynamic-based predictions of kinase-substrate interactions.

-NetPhos and PhosphoPred predict the likelihood that a particular serine, threonine, or tyrosine residue is phosphorylated by a specific kinase.

Yes, bioinformatics tools can help predict whether a protein is a substrate of a kinase. These tools use various approaches, including sequence motifs, structural features, and machine learning models trained on known kinase-substrate interactions. Here are some key methods and tools used for kinase substrate prediction:

1. Sequence-based Prediction

Kinases typically recognize short linear motifs (SLiMs) in protein sequences. Some bioinformatics tools use consensus phosphorylation site motifs to predict potential kinase-substrate interactions. Examples:

• NetPhos – Predicts phosphorylation sites based on neural networks.
• KinasePhos – Uses position-specific scoring matrices to predict kinase-specific phosphorylation sites.
• PhosphoSitePlus – A curated database of experimentally verified phosphorylation sites that can provide insights.

Thank you

Vikas Manikrao Shukre Nicolas Poirier thank you for your preciuos help. I am a molecular biologust, so it is not my specific field.

Before your answers I tried differerent bioinformatics tools, like phospsho siteplus but, if I well understood, it is based on known interaction/PTMs. So I found netphos, but it does not allow to predict if a specific kinase phosphorylates a specific substrate and it recognises on a protein sequence putative phosphorylation sites, is it correct?

Net, I found KInasephos, but it did not run.

Finally, I tried GPS 6.0, what do you think about this tool?

Yes, GPS 6.0 (Kinase-specific Phosphorylation Site Prediction) is indeed a bioinformatics tool designed to predict phosphorylation sites on proteins and their potential interactions with specific kinases. It uses a machine learning approach to identify kinase-specific phosphorylation sites based on the protein sequence.

Phosphorylation Site Prediction: GPS 6.0 predicts not only the likelihood of a phosphorylation site but also its specific kinase substrate association. It uses both sequence features and structural information to make these predictions.

Kinase Specificity: It focuses on predicting phosphorylation by specific kinases, making it very useful for researchers interested in understanding how certain kinases interact with various substrates.

Machine Learning Model: GPS 6.0 incorporates machine learning algorithms to improve prediction accuracy by learning from known kinase-substrate relationships from experimentally verified data.

High Specificity: It is tailored for kinase-specific prediction, which helps you narrow down the relevant kinases involved with a particular substrate.

Widely Used: GPS 6.0 is frequently used in the research community, which indicates its reliability and good performance.

Database-Driven: The tool is based on a large amount of experimental data, which strengthens its predictive power.

Cons:

Like any predictive model, GPS 6.0 may not always be 100% accurate. False positives or negatives can occur, especially with novel kinases or poorly characterized proteins.

Sequence Dependence: Predictions are based on sequence alone, so if the protein undergoes conformational changes that affect kinase interaction, it might not be fully captured by the tool.

Is It Reliable for Kinase-Substrate Predictions?

While GPS 6.0 is powerful and widely regarded, it’s important to complement its predictions with experimental validation (e.g., mass spectrometry, kinase assays) to confirm the predicted kinase-substrate interactions. Also, considering the complexity of phosphorylation networks, combining GPS 6.0 with other tools or databases (e.g., PhosphoSitePlus, KinasePhos) can provide a more comprehensive understanding.

Dear Elisa Caiola,Yes, your understanding is mostly correct:

• PhosphoSitePlus is a database that provides experimentally verified phosphorylation sites and known kinase-substrate relationships. It does not predict novel interactions but rather compiles existing knowledge.
• NetPhos predicts potential phosphorylation sites in a given protein sequence but does not specifically indicate which kinase is responsible for phosphorylation. It provides general predictions of phosphorylation-prone residues.
• KinasePhos is another phosphorylation site prediction tool, but if it did not run for you, there might be compatibility or server issues.
• GPS 6.0 (Group-based Prediction System) is a widely used and advanced phosphorylation site predictor. It can predict specific kinase-substrate relationships with relatively high accuracy. GPS 6.0 uses machine learning algorithms trained on known phosphorylation data and is considered reliable for computational predictions of kinase-substrate specificity.

Dear Elisa Caiola, Welcome............

Nicolas Poirier Vikas Manikrao Shukre thank you. If you are interested in helping su on this topic, please let me know. Thank you.

yes Dear