duenti/benchmarkmsagenerator

This repository contains two python scripts with the aim of generate random multiple sequence alignments that are going to be used as benchmark for locally conservation analysis methods. One of them considers only pairwise correlations between specific amino acids, the other also includes stereochemical and structural properties

benchmarkMSAgenerator

This repository contains two python scripts with the aim of generate random multiple sequence alignments that are going to be used as benchmark for locally conservation analysis methods. One of them considers only pairwise correlations between specific amino acids, the other also includes stereochemical and structural properties.

Before use it, open the script and set the global variables.

Global Variables

n_alignments: Number of alignments that are going to be generated
min_n_seqs: Minimum number of sequences in the MSA
max_n_seqs: Maximum number of sequences in the MSA
min_n_positions: Minimum number of columns
max_n_positions: Maximum number of columns
prob_high_conserved: Probability of a position be high conserved
min_prob_cons: Probability of a conserved position maintain it's conserved amino acid, otherwise an outlier residue are going to be used
min_n_main_subclass: Minimum number os type 1 subclasses
max_n_main_subclass: Maximum number of type 1 subclasses
min_n_sec_subclass: Minimum number of type 2 subclasses
max_n_sec_subclass: Maximum number of type 2 subclasses
prob_diverge: Probability of a subclass diverge (Do not share the same positions)
prob_subclass_conserved: Probability of a position be highly conserved inside a subclass
min_seqs_subclass: Minimum fraction of sequences in a subclass
prob_marginally_conservation: Probability of a conserved (or locally conserved) position be related to a marginally property. 1-p will be the probability of the position be related to an specific amino acid
prob_gap: = 0.05 #Probability of a conserved position be gap