Genetic Target selection for Authentication and Diagnosis

This repository is dedicated to genetic target selection and analysis for product authentication and diagnosis in a broader context. It also provides an R module focused on High Resolution Melting (HRM) analysis as an approach to contrast or diagnose samples.

Objectives

• MODULE 1 (Python Module to Genetic Target Selection): The main code used in this module is located in the directory target_selection/compartative_genomics/singletons_to_primers.py. The documentation for this module will be described soon. Essentially, this module is responsible for the identification of genomic regions that can be used as markers for product authentication or sample diagnosis.

• MODULE 2 (High Resolution Melting (HRM) Analysis): Providing tools in R for HRM analysis, facilitating the genotyping or diagnosis of different samples based on DNA melting characteristics.

MODULE I: GENETIC TARGET SELECTION

Documentation in progress

MODULE II: TO HRM ANALYSIS

Getting started

git clone https://github.com/Donandrade/authenticationANDdiagnostics.git

This module is based on an R script (hrm_analysis/m144.R) for performing High-Resolution Melting (HRM) analysis on the first derivative of fluorescence obtained from a real-time PCR. The script processes the data, calculates dissimilarity matrices, performs hierarchical clustering based on Genotype Confidence Percentage (GCP), and generates main plots (see below).

Dependences

The script requires the following R packages. You can install them using the commands below:

install.packages("tidyr")
if (!require("processx")) install.packages("processx")
install.packages("MBmca")

# Additionally, load the necessary libraries
library(ggplot2)
library(factoextra)
library(dendextend)
library(cluster)
library(ape)
library(circlize)
library(qpcR)
library(tidyverse)
library(plotly)
library(MBmca)
library(RColorBrewer)

The melting curve graphs will be plotted with the Plotly package, which can be exported using the additional orca command line utility. First, you need to install the orca library on your operating system (for details, see the link).

Calculation of Genotype Confidence Percentage (GCP)

According to Nunziata et al (2018) "Genotype confidence percentage (GCP) of HRM curves is the most commonly used statistical transformation of Euclidean distance between HRM curves to determine whether two curves are identical or not.". A better explanation of the procedure for calculating GCP is described below:

1. Calculation of a similarity matrix ():

   The formula is:

   Explanation:

   • is a measure of similarity.
   • The base of the exponentiation is
   • The coefficient that adjusts the influence of the sum of squared differences is
   • The summation spans an index range from to :
   • The term is the square of the difference between the fluorescence values and for each index
   • The result of the summation is multiplied by ( -0.02 ) and used as the exponent for ( 1.05 ):

2. Calculation of a dissimilatiry matrix ():

   The formula is:

   Explanation:

   • is the measure of dissimilarity.
   • It is calculated as the difference between 1 and , multiplied by 100 to convert it into a percentage:
   • The higher the value of , the greater the dissimilarity between the fluorescence profiles.

The vlaues of and are calculated by the script hrm_analysis/m144.R and a dissimilarity matrix is saved in the dissimilarity.txt file.

Usage

Place your HRM data file in the data/ directory. The example file used is named M144_Raw_Data.txt. You can use the same name for your file, or if you choose a different name, make sure to update the filename on line 22 of the code (hrm_analysis/m144.R).

Run the R script to process the data, calculate dissimilarity, perform clustering, and generate plots.

Input

An example of real data to be used as input can be found in the data directory data/M144_Raw_Data.txt. The file should be in tsv format. This dataset corresponds to the first derivative of fluorescence obtained from a real-time PCR run on the LightCycler® equipment. Below is a sample format where the first column should be the melting curve temperature and the remaining columns should contain the respective fluorescence values for each sample. In the example shown here, the data is already in the first derivative of the fluorescence.

Temp	Sample1	Sample2	Sample3	Sample4	Sample5	Sample6	Sample7
59.540	1.953	1.895	1.813	1.722	1.632	2.538	2.032
59.817	1.953	1.895	1.813	1.722	1.632	2.538	2.032
60.094	2.309	2.139	2.168	2.050	1.986	2.873	2.265
60.371	2.763	2.445	2.648	2.510	2.490	3.338	2.623
60.649	3.151	2.689	3.095	2.928	2.937	3.727	2.923
60.926	3.363	2.826	3.348	3.137	3.154	3.903	3.026
61.203	3.371	2.816	3.351	3.117	3.136	3.870	2.952
61.480	3.248	2.673	3.212	2.970	2.979	3.707	2.794
61.757	3.144	2.535	3.107	2.845	2.861	3.559	2.681
...	...	...	...	...	...	...	...

Otput

• dissimilaridade.txt for the dissimilarity matrix which includes the effect of genotype confidence percentage.
• M144.pdf: Shows the hierarchical clustering results as a dendrogram. The plot illustrates how samples are grouped into clusters based on their dissimilarity scores. For now, we suggest you check this file to select the number of K-means clusters and then choose the number of clusters. In our example, we selected k=2 (see the k on line 122 of the script hrm_analysis/m144.R).
• df_M144.pdf: Displays the -dF/dT of fluorescence data against the shift temperature for each sample. This step will require future optimization, as it currently depends on user adjustments. Therefore, be attentive and choose the best temperature shift for your data.

NOTES

• Ensure that all necessary R packages are installed.
• The script assumes that the HRM data file is formatted correctly and located in the specified directory.
• Adjust the script as needed for different data formats or analysis requirements.