Finalizing Multiple Sequence Alignment

This commit is contained in:
Noah L. Schrick 2022-11-21 17:16:46 -06:00
parent bd4cfb8688
commit 056dcb299b
9 changed files with 933 additions and 1 deletions

512
.Rhistory Normal file
View File

@ -0,0 +1,512 @@
################# Max-Log-Likelihood #################
n <- length(g.breaks.clean)
kmin <- g.breaks.clean[1]
alpha.ML <- 1 + n/sum(log(g.breaks.clean)/kmin)
alpha.ML
lines(g.seq, g.seq^(-alpha.ML), col="#D35FB7", lty=4)
# Homework 4 for the University of Tulsa' s CS-7863 Network Theory Course
# Degree Distribution
# Professor: Dr. McKinney, Spring 2022
# Noah Schrick - 1492657
library(igraph)
library(igraphdata)
data(yeast)
g <- yeast
g.netname <- "Yeast"
################# Set up Work #################
g.vec <- degree(g)
g.hist <- hist(g.vec, freq=FALSE, main=paste("Histogram of the", g.netname,
" Network"))
legend("topright", c("Guess", "Poisson", "Least-Squares Fit",
"Max Log-Likelihood"), lty=c(1,2,3,4), col=c("#40B0A6",
"#006CD1", "#E66100", "#D35FB7"))
g.mean <- mean(g.vec)
g.seq <- 0:max(g.vec) # x-axis
################# Guessing Alpha #################
alpha.guess <- 1.5
lines(g.seq, g.seq^(-alpha.guess), col="#40B0A6", lty=1)
################# Poisson #################
g.pois <- dpois(g.seq, g.mean, log=F)
lines(g.seq, g.pois, col="#006CD1", lty=2)
################# Linear model: Least-Squares Fit #################
g.breaks <- g.hist$breaks[-c(1,2,3)] # remove 0
g.probs <- g.hist$density[-1] # make lengths match
# Need to clean up probabilities that are 0
nz.probs.mask <- g.probs!=0
g.breaks.clean <- g.breaks[nz.probs.mask]
g.probs.clean <- g.breaks[nz.probs.mask]
#plot(log(g.breaks.clean), log(g.probs.clean))
g.fit <- lm(log(g.probs.clean)~log(g.breaks.clean))
summary(g.fit)
alpha.LM <- coef(g.fit)[2]
lines(g.seq, g.seq^(-alpha.LM), col="#E66100", lty=3)
################# Max-Log-Likelihood #################
n <- length(g.breaks.clean)
kmin <- g.breaks.clean[1]
alpha.ML <- 1 + n/sum(log(g.breaks.clean)/kmin)
alpha.ML
lines(g.seq, g.seq^(-alpha.ML), col="#D35FB7", lty=4)
# Homework 4 for the University of Tulsa' s CS-7863 Network Theory Course
# Degree Distribution
# Professor: Dr. McKinney, Spring 2022
# Noah Schrick - 1492657
library(igraph)
library(igraphdata)
data(yeast)
g <- yeast
g.netname <- "Yeast"
################# Set up Work #################
g.vec <- degree(g)
g.hist <- hist(g.vec, freq=FALSE, main=paste("Histogram of the", g.netname,
" Network"))
legend("topright", c("Guess", "Poisson", "Least-Squares Fit",
"Max Log-Likelihood"), lty=c(1,2,3,4), col=c("#40B0A6",
"#006CD1", "#E66100", "#D35FB7"))
g.mean <- mean(g.vec)
g.seq <- 0:max(g.vec) # x-axis
################# Guessing Alpha #################
alpha.guess <- 1.5
lines(g.seq, g.seq^(-alpha.guess), col="#40B0A6", lty=1)
################# Poisson #################
g.pois <- dpois(g.seq, g.mean, log=F)
lines(g.seq, g.pois, col="#006CD1", lty=2)
################# Linear model: Least-Squares Fit #################
g.breaks <- g.hist$breaks[-c(1)] # remove 0
g.probs <- g.hist$density[-1] # make lengths match
# Need to clean up probabilities that are 0
nz.probs.mask <- g.probs!=0
g.breaks.clean <- g.breaks[nz.probs.mask]
g.probs.clean <- g.breaks[nz.probs.mask]
#plot(log(g.breaks.clean), log(g.probs.clean))
g.fit <- lm(log(g.probs.clean)~log(g.breaks.clean))
summary(g.fit)
alpha.LM <- coef(g.fit)[2]
lines(g.seq, g.seq^(-alpha.LM), col="#E66100", lty=3)
################# Max-Log-Likelihood #################
n <- length(g.breaks.clean)
kmin <- g.breaks.clean[1]
alpha.ML <- 1 + n/sum(log(g.breaks.clean)/kmin)
alpha.ML
lines(g.seq, g.seq^(-alpha.ML), col="#D35FB7", lty=4)
# Homework 4 for the University of Tulsa' s CS-7863 Network Theory Course
# Degree Distribution
# Professor: Dr. McKinney, Spring 2022
# Noah Schrick - 1492657
library(igraph)
library(igraphdata)
data(yeast)
g <- yeast
g.netname <- "Yeast"
################# Set up Work #################
g.vec <- degree(g)
g.hist <- hist(g.vec, freq=FALSE, main=paste("Histogram of the", g.netname,
" Network"))
legend("topright", c("Guess", "Poisson", "Least-Squares Fit",
"Max Log-Likelihood"), lty=c(1,2,3,4), col=c("#40B0A6",
"#006CD1", "#E66100", "#D35FB7"))
g.mean <- mean(g.vec)
g.seq <- 0:max(g.vec) # x-axis
################# Guessing Alpha #################
alpha.guess <- 1.5
lines(g.seq, g.seq^(-alpha.guess), col="#40B0A6", lty=1)
################# Poisson #################
g.pois <- dpois(g.seq, g.mean, log=F)
lines(g.seq, g.pois, col="#006CD1", lty=2)
################# Linear model: Least-Squares Fit #################
#g.breaks <- g.hist$breaks[-c(1)] # remove 0
g.breaks <- g.hist$breaks # remove 0
g.probs <- g.hist$density[-1] # make lengths match
# Need to clean up probabilities that are 0
nz.probs.mask <- g.probs!=0
g.breaks.clean <- g.breaks[nz.probs.mask]
g.probs.clean <- g.breaks[nz.probs.mask]
#plot(log(g.breaks.clean), log(g.probs.clean))
g.fit <- lm(log(g.probs.clean)~log(g.breaks.clean))
summary(g.fit)
alpha.LM <- coef(g.fit)[2]
lines(g.seq, g.seq^(-alpha.LM), col="#E66100", lty=3)
################# Max-Log-Likelihood #################
n <- length(g.breaks.clean)
kmin <- g.breaks.clean[1]
alpha.ML <- 1 + n/sum(log(g.breaks.clean)/kmin)
alpha.ML
lines(g.seq, g.seq^(-alpha.ML), col="#D35FB7", lty=4)
# Homework 4 for the University of Tulsa' s CS-7863 Network Theory Course
# Degree Distribution
# Professor: Dr. McKinney, Spring 2022
# Noah Schrick - 1492657
library(igraph)
library(igraphdata)
data(yeast)
g <- yeast
g.netname <- "Yeast"
################# Set up Work #################
g.vec <- degree(g)
g.hist <- hist(g.vec, freq=FALSE, main=paste("Histogram of the", g.netname,
" Network"))
legend("topright", c("Guess", "Poisson", "Least-Squares Fit",
"Max Log-Likelihood"), lty=c(1,2,3,4), col=c("#40B0A6",
"#006CD1", "#E66100", "#D35FB7"))
g.mean <- mean(g.vec)
g.seq <- 0:max(g.vec) # x-axis
################# Guessing Alpha #################
alpha.guess <- 1.5
lines(g.seq, g.seq^(-alpha.guess), col="#40B0A6", lty=1)
################# Poisson #################
g.pois <- dpois(g.seq, g.mean, log=F)
lines(g.seq, g.pois, col="#006CD1", lty=2)
################# Linear model: Least-Squares Fit #################
g.breaks <- g.hist$breaks[-c(1)] # remove 0
g.probs <- g.hist$density[-1] # make lengths match
# Need to clean up probabilities that are 0
nz.probs.mask <- g.probs!=0
g.breaks.clean <- g.breaks[nz.probs.mask]
g.probs.clean <- g.probs[nz.probs.mask]
#plot(log(g.breaks.clean), log(g.probs.clean))
g.fit <- lm(log(g.probs.clean)~log(g.breaks.clean))
summary(g.fit)
alpha.LM <- coef(g.fit)[2]
lines(g.seq, g.seq^(-alpha.LM), col="#E66100", lty=3)
################# Max-Log-Likelihood #################
n <- length(g.breaks.clean)
kmin <- g.breaks.clean[1]
alpha.ML <- 1 + n/sum(log(g.breaks.clean)/kmin)
alpha.ML
lines(g.seq, g.seq^(-alpha.ML), col="#D35FB7", lty=4)
alpha.LM
# Homework 4 for the University of Tulsa' s CS-7863 Network Theory Course
# Degree Distribution
# Professor: Dr. McKinney, Spring 2022
# Noah Schrick - 1492657
library(igraph)
library(igraphdata)
data(yeast)
g <- yeast
g.netname <- "Yeast"
################# Set up Work #################
g.vec <- degree(g)
g.hist <- hist(g.vec, freq=FALSE, main=paste("Histogram of the", g.netname,
" Network"))
legend("topright", c("Guess", "Poisson", "Least-Squares Fit",
"Max Log-Likelihood"), lty=c(1,2,3,4), col=c("#40B0A6",
"#006CD1", "#E66100", "#D35FB7"))
g.mean <- mean(g.vec)
g.seq <- 0:max(g.vec) # x-axis
################# Guessing Alpha #################
alpha.guess <- 1.5
lines(g.seq, g.seq^(-alpha.guess), col="#40B0A6", lty=1)
################# Poisson #################
g.pois <- dpois(g.seq, g.mean, log=F)
lines(g.seq, g.pois, col="#006CD1", lty=2)
################# Linear model: Least-Squares Fit #################
g.breaks <- g.hist$breaks[-c(1)] # remove 0
g.probs <- g.hist$density[-1] # make lengths match
# Need to clean up probabilities that are 0
nz.probs.mask <- g.probs!=0
g.breaks.clean <- g.breaks[nz.probs.mask]
g.probs.clean <- g.probs[nz.probs.mask]
#plot(log(g.breaks.clean), log(g.probs.clean))
g.fit <- lm(log(g.probs.clean)~log(g.breaks.clean))
summary(g.fit)
alpha.LM <- coef(g.fit)[2]
lines(g.seq, g.seq^(-alpha.LM), col="#E66100", lty=3)
################# Max-Log-Likelihood #################
n <- length(g.breaks.clean)
kmin <- g.breaks.clean[1]
alpha.ML <- 1 + n/sum(log(g.breaks.clean/kmin))
alpha.ML
lines(g.seq, g.seq^(-alpha.ML), col="#D35FB7", lty=4)
# Homework 4 for the University of Tulsa' s CS-7863 Network Theory Course
# Degree Distribution
# Professor: Dr. McKinney, Spring 2022
# Noah Schrick - 1492657
library(igraph)
library(igraphdata)
data(yeast)
g <- yeast
g.netname <- "Yeast"
################# Set up Work #################
g.vec <- degree(g)
g.hist <- hist(g.vec, freq=FALSE, main=paste("Histogram of the", g.netname,
" Network"))
legend("topright", c("Guess", "Poisson", "Least-Squares Fit",
"Max Log-Likelihood"), lty=c(1,2,3,4), col=c("#40B0A6",
"#006CD1", "#E66100", "#D35FB7"))
g.mean <- mean(g.vec)
g.seq <- 0:max(g.vec) # x-axis
################# Guessing Alpha #################
alpha.guess <- 1.5
lines(g.seq, g.seq^(-alpha.guess), col="#40B0A6", lty=1, lwd=5)
################# Poisson #################
g.pois <- dpois(g.seq, g.mean, log=F)
lines(g.seq, g.pois, col="#006CD1", lty=2)
################# Linear model: Least-Squares Fit #################
g.breaks <- g.hist$breaks[-c(1)] # remove 0
g.probs <- g.hist$density[-1] # make lengths match
# Need to clean up probabilities that are 0
nz.probs.mask <- g.probs!=0
g.breaks.clean <- g.breaks[nz.probs.mask]
g.probs.clean <- g.probs[nz.probs.mask]
#plot(log(g.breaks.clean), log(g.probs.clean))
g.fit <- lm(log(g.probs.clean)~log(g.breaks.clean))
summary(g.fit)
alpha.LM <- coef(g.fit)[2]
lines(g.seq, g.seq^(-alpha.LM), col="#E66100", lty=3)
################# Max-Log-Likelihood #################
n <- length(g.breaks.clean)
kmin <- g.breaks.clean[1]
alpha.ML <- 1 + n/sum(log(g.breaks.clean/kmin))
alpha.ML
lines(g.seq, g.seq^(-alpha.ML), col="#D35FB7", lty=4)
# Homework 4 for the University of Tulsa' s CS-7863 Network Theory Course
# Degree Distribution
# Professor: Dr. McKinney, Spring 2022
# Noah Schrick - 1492657
library(igraph)
library(igraphdata)
data(yeast)
g <- yeast
g.netname <- "Yeast"
################# Set up Work #################
g.vec <- degree(g)
g.hist <- hist(g.vec, freq=FALSE, main=paste("Histogram of the", g.netname,
" Network"))
legend("topright", c("Guess", "Poisson", "Least-Squares Fit",
"Max Log-Likelihood"), lty=c(1,2,3,4), col=c("#40B0A6",
"#006CD1", "#E66100", "#D35FB7"))
g.mean <- mean(g.vec)
g.seq <- 0:max(g.vec) # x-axis
################# Guessing Alpha #################
alpha.guess <- 1.5
lines(g.seq, g.seq^(-alpha.guess), col="#40B0A6", lty=1, lwd=3)
################# Poisson #################
g.pois <- dpois(g.seq, g.mean, log=F)
lines(g.seq, g.pois, col="#006CD1", lty=2)
################# Linear model: Least-Squares Fit #################
g.breaks <- g.hist$breaks[-c(1)] # remove 0
g.probs <- g.hist$density[-1] # make lengths match
# Need to clean up probabilities that are 0
nz.probs.mask <- g.probs!=0
g.breaks.clean <- g.breaks[nz.probs.mask]
g.probs.clean <- g.probs[nz.probs.mask]
#plot(log(g.breaks.clean), log(g.probs.clean))
g.fit <- lm(log(g.probs.clean)~log(g.breaks.clean))
summary(g.fit)
alpha.LM <- coef(g.fit)[2]
lines(g.seq, g.seq^(-alpha.LM), col="#E66100", lty=3)
################# Max-Log-Likelihood #################
n <- length(g.breaks.clean)
kmin <- g.breaks.clean[1]
alpha.ML <- 1 + n/sum(log(g.breaks.clean/kmin))
alpha.ML
lines(g.seq, g.seq^(-alpha.ML), col="#D35FB7", lty=4)
# Homework 4 for the University of Tulsa' s CS-7863 Network Theory Course
# Degree Distribution
# Professor: Dr. McKinney, Spring 2022
# Noah Schrick - 1492657
library(igraph)
library(igraphdata)
data(yeast)
g <- yeast
g.netname <- "Yeast"
################# Set up Work #################
g.vec <- degree(g)
g.hist <- hist(g.vec, freq=FALSE, main=paste("Histogram of the", g.netname,
" Network"))
legend("topright", c("Guess", "Poisson", "Least-Squares Fit",
"Max Log-Likelihood"), lty=c(1,2,3,4), col=c("#40B0A6",
"#006CD1", "#E66100", "#D35FB7"))
g.mean <- mean(g.vec)
g.seq <- 0:max(g.vec) # x-axis
################# Guessing Alpha #################
alpha.guess <- 1.5
lines(g.seq, g.seq^(-alpha.guess), col="#40B0A6", lty=1, lwd=3)
################# Poisson #################
g.pois <- dpois(g.seq, g.mean, log=F)
lines(g.seq, g.pois, col="#006CD1", lty=2, lwd=3)
################# Linear model: Least-Squares Fit #################
g.breaks <- g.hist$breaks[-c(1)] # remove 0
g.probs <- g.hist$density[-1] # make lengths match
# Need to clean up probabilities that are 0
nz.probs.mask <- g.probs!=0
g.breaks.clean <- g.breaks[nz.probs.mask]
g.probs.clean <- g.probs[nz.probs.mask]
#plot(log(g.breaks.clean), log(g.probs.clean))
g.fit <- lm(log(g.probs.clean)~log(g.breaks.clean))
summary(g.fit)
alpha.LM <- coef(g.fit)[2]
lines(g.seq, g.seq^(-alpha.LM), col="#E66100", lty=3, lwd=3)
################# Max-Log-Likelihood #################
n <- length(g.breaks.clean)
kmin <- g.breaks.clean[1]
alpha.ML <- 1 + n/sum(log(g.breaks.clean/kmin))
alpha.ML
lines(g.seq, g.seq^(-alpha.ML), col="#D35FB7", lty=4, lwd=3)
plot(yeast)
hist(yeast)
hist(g.vec)
g.pois
g.mean
alpha.LM
alpha.ML
degree(g)
sort(degree(g))
sort(degree(g),decreasing=FALSE)
sort(degree(g),decreasing=F)
sort(degree(g),decreasing=false)
sort(degree(g), decreasing = TRUE)
head(sort(degree(g), decreasing = TRUE))
stddev(degree(g))
sd(degree(g))
tail(sort(degree(g), decreasing = TRUE))
plot(log(g.breaks.clean), log(g.probs.clean))
# Homework 4 for the University of Tulsa' s CS-7863 Network Theory Course
# Degree Distribution
# Professor: Dr. McKinney, Spring 2022
# Noah Schrick - 1492657
library(igraph)
library(igraphdata)
data(yeast)
g <- yeast
g.netname <- "Yeast"
################# Set up Work #################
g.vec <- degree(g)
g.hist <- hist(g.vec, freq=FALSE, main=paste("Histogram of the", g.netname,
" Network"))
legend("topright", c("Guess", "Poisson", "Least-Squares Fit",
"Max Log-Likelihood"), lty=c(1,2,3,4), col=c("#40B0A6",
"#006CD1", "#E66100", "#D35FB7"))
g.mean <- mean(g.vec)
g.seq <- 0:max(g.vec) # x-axis
################# Guessing Alpha #################
alpha.guess <- 1.5
lines(g.seq, g.seq^(-alpha.guess), col="#40B0A6", lty=1, lwd=3)
################# Poisson #################
g.pois <- dpois(g.seq, g.mean, log=F)
lines(g.seq, g.pois, col="#006CD1", lty=2, lwd=3)
################# Linear model: Least-Squares Fit #################
g.breaks <- g.hist$breaks[-c(1)] # remove 0
g.probs <- g.hist$density[-1] # make lengths match
# Need to clean up probabilities that are 0
nz.probs.mask <- g.probs!=0
g.breaks.clean <- g.breaks[nz.probs.mask]
g.probs.clean <- g.probs[nz.probs.mask]
plot(log(g.breaks.clean), log(g.probs.clean))
g.fit <- lm(log(g.probs.clean)~log(g.breaks.clean))
summary(g.fit)
alpha.LM <- coef(g.fit)[2]
lines(g.seq, g.seq^(-alpha.LM), col="#E66100", lty=3, lwd=3)
################# Max-Log-Likelihood #################
n <- length(g.breaks.clean)
kmin <- g.breaks.clean[1]
alpha.ML <- 1 + n/sum(log(g.breaks.clean/kmin))
alpha.ML
lines(g.seq, g.seq^(-alpha.ML), col="#D35FB7", lty=4, lwd=3)
plot(log(g.breaks.clean), log(g.probs.clean))
g.breaks.clean <- g.breaks[nz.probs.mask]
g.probs.clean <- g.probs[nz.probs.mask]
plot(log(g.breaks.clean), log(g.probs.clean))
## Set Working Directory to file directory - RStudio approach
setwd(dirname(rstudioapi::getActiveDocumentContext()$path))
#### Part A: GenBank sequences and a multiple fasta file
if (!require("ape")) install.packages("ape")
library(ape) # needed for read.GenBank
# fetch the mtDNA sequences
mtDNA.MultiSeqs.list<-read.GenBank(c("AF011222","AF254446","X90314","AF089820",
"AF176766","AF451972", "AY079510",
"AF050738","AF176722","AF315498",
"AF176731","AF451964"), as.character=TRUE)
# look at species names
mtDNA.Species<-attr(mtDNA.MultiSeqs.list,"species")
# use species as name instead of genbank id
names(mtDNA.MultiSeqs.list)<-mtDNA.Species
names(mtDNA.MultiSeqs.list)
# need to fix some names
names(mtDNA.MultiSeqs.list)[1] <- paste("German_Neanderthal",sep="")
names(mtDNA.MultiSeqs.list)[2] <- paste("Russian_Neanderthal",sep="")
names(mtDNA.MultiSeqs.list)[3] <- paste("Human")
names(mtDNA.MultiSeqs.list)[6] <- paste("Puti_Orangutan",sep="")
names(mtDNA.MultiSeqs.list)[12] <- paste("Jari_Orangutan",sep="")
names(mtDNA.MultiSeqs.list)
# look at one of the sequences using $
mtDNA.MultiSeqs.list$Human
length(mtDNA.MultiSeqs.list$Human)
## Convert to Biostrings object for the sequences
if (!require("BiocManager")) install.packages("BiocManager")
library(BiocManager)
if (!require("Biostrings")) BiocManager::install("Biostrings")
library(Biostrings)
# loop through the list to create vector of strings for Biostrings input
Names.vec <- c() # initialize speices names string vector
Seqs.vec <- c() # initialize sequence string vector
for (mtDNA.name in names(mtDNA.MultiSeqs.list))
{
Names.vec <- c(Names.vec,mtDNA.name) # concatenate vector
Seqs.vec <-c(Seqs.vec,paste(mtDNA.MultiSeqs.list[[mtDNA.name]],collapse=""))
}
mtDNA.multSeqs.bstr <- DNAStringSet(Seqs.vec) # convert to Biostring
# name the Biostring sequences and compute stats
names(mtDNA.multSeqs.bstr) <- Names.vec # count nucs and sequence lengths
num.nts <- alphabetFrequency(mtDNA.multSeqs.bstr)[,1:4]
mtDNA.lengths <- rowSums(num.nts)
proportion.nts <- num.nts/mtDNA.lengths
num.nts
names(mtDA.multSeqs.bstr)
names(mtDNA.multSeqs.bstr)
mtDNA.multSeqs.bstr
mtDNA.multSeqs.bstr
mtDNA.multSeqs.bstr
mtDNA.multSeqs.bstr --help
print(mtDNA.multSeqs.bstr)
print(mtDNA.multSeqs.bstr, -n40)
print(mtDNA.multSeqs.bstr, -n 40)
print(mtDNA.multSeqs.bstr, n=20)
class(mtDNA.multSeqs.bstr)
print(mtDNA.multSeqs.bstr)
?print()
?print
table(mtDNA.multSeqs.bstr)
mtDNA.multSeqs.bstr
mtDNA.multSeqs.bstr$width
mtDNA.multSeqs.bstr[,1]$width
mtDNA.multSeqs.bstr[1,]$width
mtDNA.multSeqs.bstr[1]$width
mtDNA.multSeqs.bstr[1]
mtDNA.multSeqs.bstr[1]$seq
mtDNA.multSeqs.bstr[1]$width
mtDNA.multSeqs.bstr[1]$names
mtDNA.multSeqs.bstr$names
# name the Biostring sequences and compute stats
names(mtDNA.multSeqs.bstr) <- Names.vec # count nucs and sequence lengths
mtDNA.multSeqs.bstr$names
mtDNA.multSeqs.bstr$Names
mtDNA.lengths
table(mtDNA.lengths, Names.vec)
cbind(mtDNA.lengths, Names.vec)
sort(cbind(mtDNA.lengths, Names.vec))
cbind(mtDNA.lengths, Names.vec)
cbind(mtDNA.lengths, Names.vec)
table(cbind(mtDNA.lengths, Names.vec))
rbind(cbind(mtDNA.lengths, Names.vec))
sort(rbind(cbind(mtDNA.lengths, Names.vec)))
rbind(mtDNA.lengths, Names.vec)
cbind(mtDNA.lengths, Names.vec)
max(cbind(mtDNA.lengths, Names.vec))
max(cbind(mtDNA.lengths, Names.vec))[,1]
max(cbind(mtDNA.lengths, Names.vec)[,1])
max(cbind(mtDNA.lengths, Names.vec)[1,])
max(cbind(mtDNA.lengths, Names.vec)[,1])
max(cbind(mtDNA.lengths, Names.vec))
cbind(mtDNA.lengths, Names.vec)
nlengthnames <- cbind(mtDNA.lengths, Names.vec)
max(nlengthnames[,1])
nlengthnames <- cbind(mtDNA.lengths, Names.vec)
nlengthnames[which.max(nlengthnames[,1])]
idx <- which.max(nlengthnames[,1])
idx
nlengthnames[idx, idx]
nlengthnames[idx]
nlengthnames
nlengthnames[idx,]
proportion.nts <- num.nts/mtDNA.lengths
proportion.nts

BIN
MSA.png Normal file

Binary file not shown.

After

Width:  |  Height:  |  Size: 136 KiB

View File

@ -48,7 +48,7 @@ mtDNA.multSeqs.bstr <- DNAStringSet(Seqs.vec) # convert to Biostring
# name the Biostring sequences and compute stats # name the Biostring sequences and compute stats
names(mtDNA.multSeqs.bstr) <- Names.vec # count nucs and sequence lengths names(mtDNA.multSeqs.bstr) <- Names.vec # count nucs and sequence lengths
# num.nts <- alphabetFrequency(mtDNA.multSeqs.bstr)[,1:4] num.nts <- alphabetFrequency(mtDNA.multSeqs.bstr)[,1:4]
mtDNA.lengths <- rowSums(num.nts) mtDNA.lengths <- rowSums(num.nts)
proportion.nts <- num.nts/mtDNA.lengths proportion.nts <- num.nts/mtDNA.lengths
@ -56,3 +56,40 @@ proportion.nts <- num.nts/mtDNA.lengths
nlengthnames <- cbind(mtDNA.lengths, Names.vec) nlengthnames <- cbind(mtDNA.lengths, Names.vec)
idx <- which.max(nlengthnames[,1]) idx <- which.max(nlengthnames[,1])
nlengthnames[idx,] nlengthnames[idx,]
#### Part B: Multiple Sequence Alignment
if (!require("BiocManager")) install.packages("BiocManager")
library(BiocManager)
if (!require("msa")) BiocManager::install("msa")
library(msa)
mtDNA.msa <- msa(mtDNA.multSeqs.bstr,method="ClustalOmega")
msaPrettyPrint(mtDNA.msa, file="mtDNA.pdf", output="pdf", showNames="left",
showLogo="none", askForOverwrite=FALSE, verbose=TRUE )
## loop to make results data frame
num_seqs <- length(Names.vec)
# initialize data frame
align.stats.df <- data.frame(species=rep(NA,num_seqs), seqlen=rep(0,num_seqs),
numgaps=rep(0,num_seqs), nt_a=rep(NA,num_seqs),
nt_c=rep(NA,num_seqs), nt_g=rep(NA,num_seqs),
nt_t=rep(NA,num_seqs))
# DNAbin type required for dist.dna and helpful for other calculations
mtDNA.msa.DNAbin <- as.DNAbin(mtDNA.msa)
for (i in 1:num_seqs){
seq_name <- Names.vec[i]
seq.vec <- as.character(mtDNA.msa.DNAbin[i,])
num.gaps <- sum(seq.vec=="-")
prop.nt.i <- proportion.nts[i,]
align.stats.df[i,] <- c(seq_name, mtDNA.lengths[i], num.gaps,
round(prop.nt.i[1],digits=2), round(prop.nt.i[2],digits=2),
round(prop.nt.i[3],digits=2), round(prop.nt.i[4],digits=2))
}
# write to file
write.table(align.stats.df,file="alignstats.tab",sep = "\t", row.names=FALSE, quote=FALSE)
# you can use $ operator to grab a named column from a data.frame
# similar to grabbing a named variable from a list
align.stats.df$species
align.stats.df$nt_a # strings by default
as.numeric(align.stats.df$nt_a) # convert to numeric

Binary file not shown.

13
alignstats.tab Normal file
View File

@ -0,0 +1,13 @@
species seqlen numgaps nt_a nt_c nt_g nt_t
German_Neanderthal 379 81 0.31 0.32 0.13 0.24
Russian_Neanderthal 345 76 0.33 0.34 0.11 0.22
Human 360 76 0.33 0.34 0.11 0.22
Gorilla_beringei_beringei 374 96 0.29 0.36 0.13 0.23
Pan_troglodytes_troglodytes 340 105 0.32 0.36 0.1 0.22
Puti_Orangutan 354 90 0.31 0.4 0.1 0.19
Gorilla_gorilla_gorilla 367 71 0.3 0.35 0.14 0.21
Gorilla_beringei_graueri 374 105 0.28 0.35 0.14 0.23
Pan_troglodytes_schweinfurthii 339 110 0.34 0.35 0.09 0.22
Pan_troglodytes_ellioti 411 111 0.33 0.36 0.1 0.21
Pan_troglodytes_verus 339 39 0.34 0.37 0.1 0.2
Jari_Orangutan 345 111 0.32 0.39 0.1 0.19

14
mtDNA.aux Normal file
View File

@ -0,0 +1,14 @@
\relax
\savedseqlength{1}{1}{369}
\savedseqlength{1}{2}{374}
\savedseqlength{1}{3}{374}
\savedseqlength{1}{4}{354}
\savedseqlength{1}{5}{345}
\savedseqlength{1}{6}{360}
\savedseqlength{1}{7}{379}
\savedseqlength{1}{8}{345}
\savedseqlength{1}{9}{340}
\savedseqlength{1}{10}{339}
\savedseqlength{1}{11}{411}
\savedseqlength{1}{12}{339}
\gdef \@abspage@last{2}

308
mtDNA.log Normal file
View File

@ -0,0 +1,308 @@
This is pdfTeX, Version 3.141592653-2.6-1.40.24 (TeX Live 2022/Arch Linux) (preloaded format=pdflatex 2022.11.8) 21 NOV 2022 16:54
entering extended mode
restricted \write18 enabled.
%&-line parsing enabled.
**mtDNA.tex
(./mtDNA.tex
LaTeX2e <2021-11-15> patch level 1
L3 programming layer <2022-04-10>
(/usr/share/texmf-dist/tex/latex/base/article.cls
Document Class: article 2021/10/04 v1.4n Standard LaTeX document class
(/usr/share/texmf-dist/tex/latex/base/size10.clo
File: size10.clo 2021/10/04 v1.4n Standard LaTeX file (size option)
)
\c@part=\count185
\c@section=\count186
\c@subsection=\count187
\c@subsubsection=\count188
\c@paragraph=\count189
\c@subparagraph=\count190
\c@figure=\count191
\c@table=\count192
\abovecaptionskip=\skip47
\belowcaptionskip=\skip48
\bibindent=\dimen138
)
(/usr/lib/R/library/msa/tex/texshade.sty
Package: texshade 2021/04/01 LaTeX TeXshade (v1.26)
Package `texshade', Version 1.26 of 2021/04/01.
(/usr/share/texmf-dist/tex/latex/graphics/color.sty
Package: color 2021/12/07 v1.3c Standard LaTeX Color (DPC)
(/usr/share/texmf-dist/tex/latex/graphics-cfg/color.cfg
File: color.cfg 2016/01/02 v1.6 sample color configuration
)
Package color Info: Driver file: dvips.def on input line 149.
(/usr/share/texmf-dist/tex/latex/graphics-def/dvips.def
File: dvips.def 2017/06/20 v3.1d Graphics/color driver for dvips
)
(/usr/share/texmf-dist/tex/latex/graphics/dvipsnam.def
File: dvipsnam.def 2016/06/17 v3.0m Driver-dependent file (DPC,SPQR)
))
(/usr/share/texmf-dist/tex/latex/graphics/graphics.sty
Package: graphics 2021/03/04 v1.4d Standard LaTeX Graphics (DPC,SPQR)
(/usr/share/texmf-dist/tex/latex/graphics/trig.sty
Package: trig 2021/08/11 v1.11 sin cos tan (DPC)
)
(/usr/share/texmf-dist/tex/latex/graphics-cfg/graphics.cfg
File: graphics.cfg 2016/06/04 v1.11 sample graphics configuration
)
Package graphics Info: Driver file: pdftex.def on input line 107.
(/usr/share/texmf-dist/tex/latex/graphics-def/pdftex.def
File: pdftex.def 2020/10/05 v1.2a Graphics/color driver for pdftex
))
\structurefile=\read2
\featurefile=\write3
\alignfile=\read3
\sublogofile=\read4
\exp@rtfile=\write4
\exp@rt@chimerafile=\write5
(/usr/share/texmf-dist/tex/latex/amsfonts/amssymb.sty
Package: amssymb 2013/01/14 v3.01 AMS font symbols
(/usr/share/texmf-dist/tex/latex/amsfonts/amsfonts.sty
Package: amsfonts 2013/01/14 v3.01 Basic AMSFonts support
\@emptytoks=\toks16
\symAMSa=\mathgroup4
\symAMSb=\mathgroup5
LaTeX Font Info: Redeclaring math symbol \hbar on input line 98.
LaTeX Font Info: Overwriting math alphabet `\mathfrak' in version `bold'
(Font) U/euf/m/n --> U/euf/b/n on input line 106.
))
\symalphahelix=\mathgroup6
\loopcount=\count193
\innerloopcount=\count194
\outerloopcount=\count195
\seq@count=\count196
\killseq@count=\count197
\seq@percent=\count198
\res@count=\count199
\seq@pointer=\count266
\pos@count=\count267
\res@perline=\count268
\end@count=\count269
\cons@count=\count270
\total@count=\count271
\temp@count=\count272
\triple@count=\count273
\temp@@count=\count274
\pos@sum=\count275
\box@width=\skip49
\name@width=\skip50
\box@depth=\skip51
\width@tmp=\skip52
\box@height=\skip53
\number@width=\skip54
\line@stretch=\skip55
\center@fill=\skip56
\arrow@width=\skip57
\arrow@height=\skip58
\rule@thick=\skip59
\arrow@thick=\skip60
\logo@height=\skip61
\equal@width=\skip62
\equal@tmp=\skip63
\equal@height=\skip64
\temp@@length=\skip65
\vspace@legend=\skip66
\hspace@legend=\skip67
\bar@length=\skip68
Package color Info: Redefining color LightGray on input line 1716.
Package color Info: Redefining color LightLightGray on input line 1817.
Package color Info: Redefining color LightLightLightGray on input line 1919.
)
(/usr/share/texmf-dist/tex/latex/l3backend/l3backend-pdftex.def
File: l3backend-pdftex.def 2022-04-14 L3 backend support: PDF output (pdfTeX)
\l__color_backend_stack_int=\count276
\l__pdf_internal_box=\box50
)
No file mtDNA.aux.
\openout1 = `mtDNA.aux'.
LaTeX Font Info: Checking defaults for OML/cmm/m/it on input line 24.
LaTeX Font Info: ... okay on input line 24.
LaTeX Font Info: Checking defaults for OMS/cmsy/m/n on input line 24.
LaTeX Font Info: ... okay on input line 24.
LaTeX Font Info: Checking defaults for OT1/cmr/m/n on input line 24.
LaTeX Font Info: ... okay on input line 24.
LaTeX Font Info: Checking defaults for T1/cmr/m/n on input line 24.
LaTeX Font Info: ... okay on input line 24.
LaTeX Font Info: Checking defaults for TS1/cmr/m/n on input line 24.
LaTeX Font Info: ... okay on input line 24.
LaTeX Font Info: Checking defaults for OMX/cmex/m/n on input line 24.
LaTeX Font Info: ... okay on input line 24.
LaTeX Font Info: Checking defaults for U/cmr/m/n on input line 24.
LaTeX Font Info: ... okay on input line 24.
(/usr/share/texmf-dist/tex/context/base/mkii/supp-pdf.mkii
[Loading MPS to PDF converter (version 2006.09.02).]
\scratchcounter=\count277
\scratchdimen=\dimen139
\scratchbox=\box51
\nofMPsegments=\count278
\nofMParguments=\count279
\everyMPshowfont=\toks17
\MPscratchCnt=\count280
\MPscratchDim=\dimen140
\MPnumerator=\count281
\makeMPintoPDFobject=\count282
\everyMPtoPDFconversion=\toks18
) (/usr/share/texmf-dist/tex/latex/epstopdf-pkg/epstopdf-base.sty
Package: epstopdf-base 2020-01-24 v2.11 Base part for package epstopdf
Package epstopdf-base Info: Redefining graphics rule for `.eps' on input line 4
85.
(/usr/share/texmf-dist/tex/latex/latexconfig/epstopdf-sys.cfg
File: epstopdf-sys.cfg 2010/07/13 v1.3 Configuration of (r)epstopdf for TeX Liv
e
))
(/tmp/Rtmp9tSm8e/seq57d10064df.fasta:
Runaway argument?
! Paragraph ended before \inf@@get was complete.
<to be read again>
\par
l.25 ...hade}{/tmp/Rtmp9tSm8e/seq57d10064df.fasta}
I suspect you've forgotten a `}', causing me to apply this
control sequence to too much text. How can we recover?
My plan is to forget the whole thing and hope for the best.
! Misplaced alignment tab character &.
\msfline ->\par &
& & & @
l.25 ...hade}{/tmp/Rtmp9tSm8e/seq57d10064df.fasta}
I can't figure out why you would want to use a tab mark
here. If you just want an ampersand, the remedy is
simple: Just type `I\&' now. But if some right brace
up above has ended a previous alignment prematurely,
you're probably due for more error messages, and you
might try typing `S' now just to see what is salvageable.
! Misplaced alignment tab character &.
\msfline ->\par & &
& & @
l.25 ...hade}{/tmp/Rtmp9tSm8e/seq57d10064df.fasta}
I can't figure out why you would want to use a tab mark
here. If you just want an ampersand, the remedy is
simple: Just type `I\&' now. But if some right brace
up above has ended a previous alignment prematurely,
you're probably due for more error messages, and you
might try typing `S' now just to see what is salvageable.
! Misplaced alignment tab character &.
\msfline ->\par & & &
& @
l.25 ...hade}{/tmp/Rtmp9tSm8e/seq57d10064df.fasta}
I can't figure out why you would want to use a tab mark
here. If you just want an ampersand, the remedy is
simple: Just type `I\&' now. But if some right brace
up above has ended a previous alignment prematurely,
you're probably due for more error messages, and you
might try typing `S' now just to see what is salvageable.
! Misplaced alignment tab character &.
\msfline ->\par & & & &
@
l.25 ...hade}{/tmp/Rtmp9tSm8e/seq57d10064df.fasta}
I can't figure out why you would want to use a tab mark
here. If you just want an ampersand, the remedy is
simple: Just type `I\&' now. But if some right brace
up above has ended a previous alignment prematurely,
you're probably due for more error messages, and you
might try typing `S' now just to see what is salvageable.
Runaway argument?
! Paragraph ended before \check@letter was complete.
<to be read again>
\par
l.25 ...hade}{/tmp/Rtmp9tSm8e/seq57d10064df.fasta}
I suspect you've forgotten a `}', causing me to apply this
control sequence to too much text. How can we recover?
My plan is to forget the whole thing and hope for the best.
Runaway argument?
! Paragraph ended before \firstchar@get was complete.
<to be read again>
\par
l.25 ...hade}{/tmp/Rtmp9tSm8e/seq57d10064df.fasta}
I suspect you've forgotten a `}', causing me to apply this
control sequence to too much text. How can we recover?
My plan is to forget the whole thing and hope for the best.
Runaway argument?
! Paragraph ended before \firstchar@get was complete.
<to be read again>
\par
l.47 \end{texshade}
I suspect you've forgotten a `}', causing me to apply this
control sequence to too much text. How can we recover?
My plan is to forget the whole thing and hope for the best.
Runaway argument?
! Paragraph ended before \res@get was complete.
<to be read again>
\par
l.47 \end{texshade}
I suspect you've forgotten a `}', causing me to apply this
control sequence to too much text. How can we recover?
My plan is to forget the whole thing and hope for the best.
! Misplaced alignment tab character &.
\seq@line ->\par &
@
l.47 \end{texshade}
I can't figure out why you would want to use a tab mark
here. If you just want an ampersand, the remedy is
simple: Just type `I\&' now. But if some right brace
up above has ended a previous alignment prematurely,
you're probably due for more error messages, and you
might try typing `S' now just to see what is salvageable.
. . . . . [1
Non-PDF special ignored!
<special> papersize=794.96999pt,614.295pt
{/var/lib/texmf/fonts/map/pdftex/updmap/pdftex.map}] )
LaTeX Font Info: Trying to load font information for U+msa on input line 47.
(/usr/share/texmf-dist/tex/latex/amsfonts/umsa.fd
File: umsa.fd 2013/01/14 v3.01 AMS symbols A
)
LaTeX Font Info: Trying to load font information for U+msb on input line 47.
(/usr/share/texmf-dist/tex/latex/amsfonts/umsb.fd
File: umsb.fd 2013/01/14 v3.01 AMS symbols B
) [2] (./mtDNA.aux) )
Here is how much of TeX's memory you used:
8179 strings out of 478238
97095 string characters out of 5850456
1334999 words of memory out of 5000000
26433 multiletter control sequences out of 15000+600000
471599 words of font info for 39 fonts, out of 8000000 for 9000
1141 hyphenation exceptions out of 8191
541i,6n,65p,182b,4214s stack positions out of 5000i,500n,10000p,200000b,80000s
</usr/share/texmf-dist/fonts/type1/public/amsfonts/cm/cmr1
0.pfb></usr/share/texmf-dist/fonts/type1/public/amsfonts/cm/cmsy10.pfb></usr/sh
are/texmf-dist/fonts/type1/public/amsfonts/cm/cmtt10.pfb>
Output written on mtDNA.pdf (2 pages, 86917 bytes).
PDF statistics:
26 PDF objects out of 1000 (max. 8388607)
15 compressed objects within 1 object stream
0 named destinations out of 1000 (max. 500000)
1 words of extra memory for PDF output out of 10000 (max. 10000000)

BIN
mtDNA.pdf Normal file

Binary file not shown.

48
mtDNA.tex Normal file
View File

@ -0,0 +1,48 @@
\documentclass[10pt]{article}
\usepackage{texshade}
\headheight=0pt
\headsep=0pt
\hoffset=0pt
\voffset=0pt
\paperwidth=11in
\paperheight=8.5in
\ifx\pdfoutput\undefined
\relax
\else
\pdfpagewidth=\paperwidth
\pdfpageheight=\paperheight
\fi
\oddsidemargin=-0.9in
\topmargin=-0.7in
\textwidth=10.8in
\textheight=7.9in
\pagestyle{empty}
\begin{document}
\begin{texshade}{/tmp/Rtmp9tSm8e/seq57d10064df.fasta}
\seqtype{N}
\shadingmode{identical}
\threshold{50}
\showconsensus[ColdHot]{bottom}
\shadingcolors{blues}
\hidelogoscale
\shownames{left}
\nameseq{1}{Gorilla gorilla gorilla}
\nameseq{2}{Gorilla beringei beringei}
\nameseq{3}{Gorilla beringei graueri}
\nameseq{4}{Puti Orangutan}
\nameseq{5}{Jari Orangutan}
\nameseq{6}{Human}
\nameseq{7}{German Neanderthal}
\nameseq{8}{Russian Neanderthal}
\nameseq{9}{Pan troglodytes troglodytes}
\nameseq{10}{Pan troglodytes schweinfurthii}
\nameseq{11}{Pan troglodytes ellioti}
\nameseq{12}{Pan troglodytes verus}
\shownumbering{right}
\showlegend
\end{texshade}
\end{document}