PsychoProt: Physical CHemistry Of Protein variability

A tool to fit amino acid variability data to physicochemical descriptors (know more...)

Inputs

Help on the data format...

Here's a tool to build the data table from an alignment...

Paste below or Load from: Sample data sets:

Controls and Execution

Set parameters and click "Analyze"

Data includes:

k* values (= effective number of tolerated amino acids) (guessed if unavailable)

Additional column with numeric data to plot

Criteria for Fit Selection:

Filter by p-value < Rec. 0.01 - 0.05
Bonferroni correction Recommended

Filter also by Root Mean Square
Error and Pearson coefficient
|r| > For ex. 0.67082 (r² > 0.45)
RMSE < For ex. 1

Test quadratic fits even if linear fits worked

Tested descriptors:

Test for... (in this order)
* Set 1: Physicochemical descriptors
* Set 2: Metabolic descriptors
* Set 3: Extended descriptors and combinations
* Set 4: Discrete descriptors

Replacement of best fitted descriptors
p-value to change fit: Rec. 0.01 - 0.05

Core Results (Fits)

All fits that satisfy r², RMSD and p-value filters
(p-values altered by the Bonferroni correction if that was selected)
(Click to see fits)

Best fit that satisfies r² and RMSD
(Click to see fits)

Scatter plots of data and fitting curves

|||| ||||

Results mapped on protein sequence

Map best fit for each residue and compare to other data

Select all descriptors that include:

Show: all positive negative

Results by property

Plots summarizing how many times each descriptor was selected, all times (left) and as the best one (right)

Results by amino acid type

Compare ΔΔG averages throughout the full dataset (top) and when the nonnative amino acid (bottom)

Map results to 3D structure

Start here --> PDB ID: Chain:
Hint: Use a PDB whose sequence is as close as possible to your reference sequence

All fits
(multi-selection)

Trends: all (+) (-)

Best fits
(multi-selection)

Trends: all (+) (-)

Shown residues:
(click to see as CPK)

Controls:

Spin

Color by k* (slow)
Log color
Color zoom:

Highlight residues

Select all fits that include:

Select best fits that include: