Code and data for analyzing results from PEDAL (Parallel Equilibrium Dialysis for Affinity Learning) experiments by Zhao and Grigoryan.
The data and code in this repository are meant to reproduce the results of our study for establishing and validating PEDAL on the example of measuring a large number of PDZ-peptide interactions. The code is also meant to serve as demonstration of how PEDAL data can be analyzed to produce both Kd estimates and associated uncertainties.
ms_allData3.xlsx -- mass spectrometry data (including peptide sequences and peak intencities) obtained as part of our study of PDZ-peptide interactions using PEDAL. The file contains five independent datasets (different peptide libraries or different experimental settings). This is the default dataset read by analyzeMS.m.
Kd_FP_sum.xlsx -- peptide affinity validation data obtained using using Fluorescence Polarization (FP) in low throughput.
analyzeMS.m -- the main script for analyzing PEDAL data. It reads a data file (by default, ms_allData3.xlsx) and output a CSV file containing Kd, estimated error, observed apparent error, number of occurrences, alpha and cross-correlation value of each unique peptide. The script also reads Kd_FP_sum.xlsx and uses plotFpMsKd.m to plot the affinities from PEDAL assay against affinities from FP assay for a group of peptides
best.txt -- the input file for cluser_seqs.m
cluster_seqs.m -- this script reads best.txt and perform a t-SNE analysis of all the peptides included.
'results' folder contains the output result files for all binders and non-binders for each library and 4-library combined (except for the low domain experiment).