TRFLP Analysis
Microbial community analysis often relies on techniques that allow for the comparison of communities by fingerprinting technologies such as DGGE (denaturing gradient gel electrophoresis) or TRFLP (terminal fragment length polymorphisms). Once these data are obtained it becomes a challenge to analyze the distribution of microbes in these samples. A convenient way to asses similarities among samples is by means of cluster analysis. Cluster analysis requires that the data are reduced to a percent similarity matrix. For this purpose I developed Matlab code that can import a large number of TRFLP patterns and perform pair-wise comparisons.
------> DOWNLOAD CODE HERE <------
When analyzing TRFLP patterns many parameters need to be considered. The parameters the code will take into account are as follows:
 |
|---|
alpha -- finds highest peak in a TRFLP pattern (max_peak_height) and then sets to ignore all peaks that are less than alpha*max_peak_height.
beta – width of search window; searches for peaks in other TRFLP patterns that are the same size +-nucleotides in size.
gamma – minimum peak height in Fluorescence units (FU) ; ignore all peaks that are smaller.
delta – allows you to only consider those peaks that account for delta*100% of the combined area of all peaks found in a TRFLP pattern.
epsilon – minimum fragment length; allows you to consider only those peaks that are longer than epsilon nucleotides.
Once you run the code you will get a series of output files:
–alternate_distance_matrix.meg
–sorenson_distance_matrix.meg
Tab delineated data matrices
–number_of_shared_peaks.txt
–number_of_unique_peaks.txt
–significant_peaks.txt
I’ve included some sample files. The TRFLP patterns that are included
are from replicate enrichments of soil on acetate and leucine.
Distance matrices that can be analyzed using MEGA (http://www.megasoftware.net/).
Mathworks sells a bioinformatics toolbox that let’s you draw Neighbor Joining trees, but MEGA is free and widely used.
To run the code open "analysis_one_enzyme_V9_2.m"
Feel free to email me, if the code needs explanation or if you find a bug.