dist: Calculate Distances between Sets of Sequences

In mhahsler/rMSA: Interface for Popular Multiple Sequence Alignment Tools

Calculate Distances between Sets of Sequences

Description

Implements different methods to calculate distance between sets of sequences based on k-mer distribution, edit distance/alignment or evolutionary distance.

Usage

# k-mer-based methods
distFFP(x, k=3, method="JSD", normalize=TRUE)
distCV(x, k=3)
distNSV(x, k=3, method="Manhattan", normalize=FALSE)
distKMer(x, k=3)
distSimRank(x, k=7)

# edit distance/alignment
distEdit(x)
distAlignment(x, substitutionMatrix=NULL, ...)

# evolutionary distance
distApe(x, model="K80" ,...)

Arguments

x	an object of class XStringSet containing the sequences. For distApe, x needs to be a multiple sequence alignment.
k	size of used k-mers.
method	metric used to calculate the dissimilarity between two k-mer frequency distributions.
substitutionMatrix	matrix with substitution scores (defaults to a matrix with match=1, mismatch=0)
normalize	normalize the k-mer frequencies by the total number of k-mers in the sequence.
model	evolutionary model used.
...	further arguments passed on.

Details

• Feature frequency profile (distFFP): A FFP is the normalized (by the number of k-mers in the sequence) count of each possible k-mer in a sequence. The distance is defined as the Jensen-Shannon divergence (JSD) between FFPs (Sims and Kim, 2011).

• Composition Vector (distCV): A CV is a vector with the frequencies of each k-mer in the sequency minus the expected frequency of random background of neutral mutations obtained from a Markov Model. The cosine distance is used between CVs. (Qi et al, 2007).

• Numerical Summarization Vector (distNSV): An NSV is frequency distribution of all possible k-mers in a sequence. The Manhattan distance is used between NSVs (Nagar and Hahsler, 2013).

• Distance between sets of k-mers (distkMer): Each sequence is represented as a set of k-mers. The Jaccard (binary) distance is used between sets (number of unique shared k-mers over the total number of unique k-mers in both sequences).

• Distance based on SimRank (distSimRank): 1-simRank (see simRank).

• Edit (Levenshtein) Distance (distEdit): Edit distance between sequences.

• Distance based on alignment score (distAlignment): see stringDist in Biostrings.

• Evolutionary distances (distApe): see dist.dna in ape.

Value

A dist object.

Author(s)

Michael Hahsler

References

Sims, GE; Kim, SH (2011 May 17). "Whole-genome phylogeny of Escherichia coli/Shigella group by feature frequency profiles (FFPs).". Proceedings of the National Academy of Sciences of the United States of America 108 (20): 8329-34. PMID 21536867.

Gao, L; Qi, J (2007 Mar 15). "Whole genome molecular phylogeny of large dsDNA viruses using composition vector method.". BMC evolutionary biology 7: 41. PMID 17359548.

Qi J, Wang B, Hao B: Whole Proteome Prokaryote Phylogeny without Sequence Alignment: A K-String Composition Approach. Journal of Molecular Evolution 2004, 58:1-11.

Anurag Nagar; Michael Hahsler (2013). "Fast discovery and visualization of conserved regions in DNA sequences using quasi-alignment." BMC Bioinformatics, 14(Suppl. 11), 2013

Examples

s <- mutations(random_sequences(100), 100)
s

### calculate NSV distance
dNSV <- distNSV(s)

### relationship with edit distance
dEdit <- distEdit(s)

df <- data.frame(dNSV=as.vector(dNSV), dEdit=as.vector(dEdit))
plot(sapply(df, jitter), cex=.1)
### add lower bound (2*k, for Manhattan distance)
abline(0,1/(2*3), col="red", lwd=2)
### add regression line
abline(lm(dEdit~dNSV, data=df), col="blue", lwd=2)

### check correlation
cor(dNSV,dEdit)

mhahsler/rMSA documentation built on May 24, 2024, 3:36 p.m.