In general, this tool generates a graph indicating which residues of the analysed sequence are likely to form alpha helices, beta sheets, and coil regions. This makes precise and reliable characterization essential, as it allows for the creation of identifiable regular patterns associated primarily with the interactions and function of proteins.
PsiPred - Protein Analysis Workbench
PsiPred is an advanced designed for protein prediction and annotation, with an emphasis on structuring protein data. Among its tools the following stand out:
PsiPred 4 predicts the structure of proteins with deep neural networks and PSI-BLAST data.
Input
Submit the amino acid sequence of interest in FASTA format.
Algorithm
Uses protein databases to find homologies by perfoming a multiple sequence alignment, comparing the submitted protein sequence with known protein sequences, in addition to artificial neural networks to predict the secondary structure of the protein;
There are three main types of secondary structure you try to predict:
Alpha-helices: helical structures in the protein;
Beta-sheets: sheet-shaped structures organized in layers;
Random structures (coil): regions of the protein that do not have a specific secondary structure.
Results
Generates a predicton about which regions of the protein will form these secondary structure. The results are presented in the form of graphs and representation that show the probability of each structure constituting a part of the amino acid sequence.
DISOPred 3 identifies and annotates disordered regions (IDRs) using independent predictors and a result combination module.
Input
Receives the same amino acid sequence as PsiPred4 (FASTA sequence)
Algorithm
Extracts features from the amino acid sequence, such as amino acid composition, sequence patterns and other biochemical properties that may be associated with disordered regions.
It uses artificial neural networks to predict strucutral disorder, which contains known protein data to identify patterns associated with disordered regions.
Making predictions about which parts of the amino acid sequence are prone to forming disordered regions.
Results
Are generated in graphs or maps that show the probability of disorder at each position in the sequence. This can be visualized as a probability line along the sequence, where regions with a high probability of disorder are highlighted.
PsiPred and its tools are fundamental for the structural and functional analysis of proteins. PsiPred 4 predicts the secondary structure of new amino acid sequences based on evolutionarily related proteins, while DisoPred 3 identifies disordered regions, helping to detect segments without fixed structure. Together, they provide valuable information about the characteristics of proteins.
Online program that assigns a unique score to protein sequences, allowing you to predict whether the protein is intrinsically disordered or not.
Input
The sequence must be provided in FASTA format;
Algorithm
Uses physicochemical characteristics of the amino acid sequence to calculate a "folding index".
Fold index: generates a numerical value that reflects the protein's overall propensity to form an ordered structure. The positive value indicates a greater probability of forming an ordered strucutre, while the negative value suggests a tendency towards disordered regions.
Results
Presented in graphs or maps that show rhe folding rate along the amino acid sequence. Making it possible to visualize the parts of the protein that are most likely to form ordered or disordered structures.
FoldIndex estimates the propensity of a protein sequence to form an ordered or disordered structure. Based on sequence characteristics. Helps predict whether a protein will have a stable structural conformation or be flexible.
Tool that predicts intrinsically disordered regions (IDRs) in proteins using machine learning methods to assess the probability of disorder in different parts of the protein sequence.
Input
Submit the amino acid sequence of interest in FASTA format.
Algorithm
Analyzes the amino acid sequence, calculating the characteristics related to the disorder.
Amino acid composition: presence of certain amino acids, they are related to regions of disorder;
Physicochemical properties: polarity and charge of amino acids are also considere.
It uses artificial neural networks and statistical models to predict IDRs. Models are trained on databases of known proteins.
The software calculates the probability that each residue in the sequence is part of a disordered region
Results
Graphs or maps are generated that show the probability of disorder throughout the sequence. Helping to visualize the parts of the protein that are most likely to be disordered.
IuPred uses an energy model to identify segments that do not have a fixed structure, helping researchers better understand the functional and structural characteristics of proteins.