I am still new to R programming and I just have no idea how to write this same code below from python to R.
human_data is dataframe from CSV file. the word includes sequence of letters. Basically, I want to convert my 'word' column sequence of string into all possible k-mer words of length 6.
def getKmers(sequence, size=6):
return [sequence[x:x+size] for x in range(len(sequence) - size + 1)]
human_data['words'] = human_data.apply(lambda x: getKmers(x['sequence']), axis=1)
You could use the library quanteda too, in order to compute the k-mers (k-grams), the following code shows an example:
library(quanteda)
k = 6 # 6-mers
human_data = data.frame(sequence=c('abcdefghijkl', 'xxxxyyxxyzz'))
human_data$words <- apply(human_data, 1,
function(x) char_ngrams(unlist(tokens(x['sequence'],
'character')), n=k, concatenator = ''))
human_data
# sequence words
#1 abcdefghijkl abcdef, bcdefg, cdefgh, defghi, efghij, fghijk, ghijkl
#2 xxxxyyxxyzz xxxxyy, xxxyyx, xxyyxx, xyyxxy, yyxxyz, yxxyzz
I hope this helps, using R basic commands:
df = data.frame(words=c('asfdklajsjahk', 'dkajsadjkfggfh', 'kfjlhdaDDDhlw'))
getKmers = function(sequence, size=6) {
kmers = c()
for (x in 1:(nchar(sequence) - size + 1)) {
kmers = c(kmers, substr(sequence, x, x+size-1))
}
return(kmers)
}
sapply(df$words, getKmers)
Related
I have seen a number of questions similar to this but my cluster labels consist of sentence embeddings, thus a better question may be how do I get text values from the sentence embeddings?
How can I get from my sentence embeddings to print a text output?
umap_embeddings = umap.UMAP(n_neighbors=50,
n_components=5,
metric='cosine').fit_transform(embeddings)
cluster = hdbscan.HDBSCAN(min_cluster_size=3,
metric='euclidean',
cluster_selection_method='eom').fit(umap_embeddings)
# Prepare data
umap_data = umap.UMAP(n_neighbors=15, n_components=2, min_dist=0.0, metric='cosine', random_state=24).fit_transform(embeddings)
result = pd.DataFrame(umap_data, columns=['x', 'y'])
result['labels'] = cluster.labels_
# Visualize clusters
fig, ax = plt.subplots(figsize=(20, 10))
outliers = result.loc[result.labels == -1, :]
clustered = result.loc[result.labels != -1, :]
plt.scatter(outliers.x, outliers.y, color='#202020', s=25)
plt.scatter(clustered.x, clustered.y, c=clustered.labels, s=25, cmap='hsv_r'
)
some previous answers have suggested;
textdata_with_label_113 = textData[clusterer.labels_ == 113]
However, this returns the embedded value oppoesd to the text value.
With more time on the problem I realised that the embeddings are in the same sequence as the original DF.
therefore you can work back quite easily.
lbls=[]
#seperate the clustered labels into seperate lists (0,1,2,3)
for x in range(len(clustered.labels)):
lbls.append(clustered[clustered.labels == x])
df_desc=[]
# extract the rows from the data frame using the lbls list and use column 6 only in my case
for x in range(len(lbls)):
df_desc.append(df.iloc[lbls[x].index,5])
for i in range(4):
txt = "Cluster {number}"
print(txt.format(number = i))
print(df_desc[i])
I am trying to visualize sequences of events by using Sankey diagrams.
I have a set of event (Event1 to Event16) over sequences of different length.
The steps of the sequences are noted by T0, T0 - 1, T0 - 2 ...
The width of the flow is corresponding to the frequency rate of the sequences.
I would like that all the nodes corresponding to a given step to be aligned vertically.
By using the GoogleVis package I succeed to obtain the following :
Sankey with GoogleVis
As you can see some events T0-1, T0-2 and T0-3... are on the far right, instead of with the others of their time step.
It seems to be due to the fact that it is not possible to have nodes whithout children...
Do you know a way to have hierarchize nodes or/and nodes whithout children, for GoogleVis ?
If not, do you know another R package which could allow to have these characteristics for interactive plots ?
My R code is bellow. The main variable containing the sequences is a list of list, see picture.
Data containing sequences
My code :
# Package
library(googleVis)
library(dplyr)
library(reshape2)
library(tidyverse)
# Load
load("SeqCh")
# Loop -------------------------------------------------------------
# Inits
From = c()
To = c()
Freq = c()
Target = SeqCh
# Get maximum length of sequence
maxls = 0
for (kk in 1:length(Target)){
temp = length(Target[[kk]])
if (temp > maxls){
maxls = temp
}
}
# Loop on length of sequences
for (zz in 2:maxls){
# Prefix to add to manage same event repeated
if (zz == 2){
SufixFrom = "(T0)"
SufixTo = "(T0 - 1)"
} else {
SufixFrom = paste("(T0 - ", as.character(zz-2), ")", sep = "")
SufixTo = paste("(T0 - ", as.character(zz-1), ")", sep = "")
}
# Message
cat("\n")
print(paste(" Processing events from ", SufixFrom, " to ", SufixTo))
# Loop on Target
ind = lapply(Target, function(x) length(x) == zz)
TargetSub = Target[unlist(ind)]
FreqSub = Support[unlist(ind)]
for (jj in 1:length(TargetSub)){
temp = TargetSub[[jj]]
TempFrom = paste(temp[zz-1], SufixFrom, sep = " ")
TempTo = paste(temp[zz], SufixTo, sep = " ")
From = c(From, TempFrom)
To = c(To, TempTo)
Freq = c(Freq, FreqSub[jj])
}
} # end for loop on length of sequences
# All in same variable
Flows = data.frame("From" = From, "To" = To, "Occurence_Frequency" = Freq, stringsAsFactors = FALSE)
# Plot --------------------------------------------------------------------
plot(gvisSankey(Flows, from='From', to='To', weight="Occurence_Frequency",
options=list(height=900, width=1800, sankey="{link:{color:{fill:'lightblue'}}}")))
Thanks, Romain.
I am trying to efficiently map exact peptides (short sequences of amino acids in the 26 character alphabet A-Z1) to proteins (longer sequences of the same alphabet). The most efficient way to do this I'm aware of is an Aho-Corasick trie (where peptides are the keywords). Unfortunately I can't find a version of AC in R that will work with a non-nucleotide alphabet (Biostrings' PDict and Starr's match_ac are both hard-coded for DNA).
As a crutch I've been trying to parallelize a basic grep approach. But I'm having trouble figuring out a way to do so without incurring significant IO overhead. Here is a brief example:
peptides = c("FSSSGGGGGGGR","GAHLQGGAK","GGSGGSYGGGGSGGGYGGGSGSR","IISNASCTTNCLAPLAK")
if (!exists("proteins"))
{
biocLite("biomaRt", ask=F, suppressUpdates=T, suppressAutoUpdate=T)
library(biomaRt)
ensembl = useMart("ensembl",dataset="hsapiens_gene_ensembl")
proteins = getBM(attributes=c('peptide', 'refseq_peptide'), filters='refseq_peptide', values=c("NP_000217", "NP_001276675"), mart=ensembl)
row.names(proteins) = proteins$refseq_peptide
}
library(snowfall)
library(Biostrings)
library(plyr)
sfInit(parallel=T, cpus=detectCores()-1)
allPeptideInstances = NULL
i=1
increment=100
count=nrow(proteins)
while(T)
{
print(paste(i, min(count, i+increment), sep=":"))
text_source = proteins[i:min(count, i+increment),]
text = text_source$peptide
#peptideInstances = sapply(peptides, regexpr, text, fixed=T, useBytes=T)
peptideInstances = sfSapply(peptides, regexpr, text, fixed=T, useBytes=T)
dimnames(peptideInstances) = list(text_source$refseq_peptide, colnames(peptideInstances))
sparsePeptideInstances = alply(peptideInstances, 2, .fun = function(x) {x[x > 0]}, .dims = T)
allPeptideInstances = c(allPeptideInstances, sparsePeptideInstances, recursive=T)
if (i==count | nrow(text_source) < increment)
break
i = i+increment
}
sfStop()
There are a few issues here:
peptideInstances here is a dense matrix, so
returning it from each worker is very verbose. I have broken it up
into blocks so that I'm not dealing with a 40,000 (proteins) x 60,000
(peptides) matrix.
Parallelizing over peptides, when it would make
more sense to parallelize over the proteins because they're bigger.
But I got frustrated with trying to do it by protein because:
This code breaks if there is only one protein in text_source.
Alternatively, if anyone is aware of a better solution in R, I'm happy to use that. I've spent enough time on this I probably would have been better served implementing Aho-Corasick.
1 Some of those are ambiguity codes, but for simplicity, ignore that.
I learned Rcpp and implemented an Aho-Corasick myself. Now CRAN has a good general purpose multiple-keyword search package.
Here are some usage examples:
listEquals = function(a, b) { is.null(unlist(a)) && is.null(unlist(b)) || !is.null(a) && !is.null(b) && all(unlist(a) == unlist(b)) }
# simple search of multiple keywords in a single text
keywords = c("Abra", "cadabra", "is", "the", "Magic", "Word")
oneSearch = AhoCorasickSearch(keywords, "Is Abracadabra the Magic Word?")
stopifnot(listEquals(oneSearch[[1]][[1]], list(keyword="Abra", offset=4)))
stopifnot(listEquals(oneSearch[[1]][[2]], list(keyword="cadabra", offset=8)))
stopifnot(listEquals(oneSearch[[1]][[3]], list(keyword="the", offset=16)))
stopifnot(listEquals(oneSearch[[1]][[4]], list(keyword="Magic", offset=20)))
stopifnot(listEquals(oneSearch[[1]][[5]], list(keyword="Word", offset=26)))
# search a list of lists
# * sublists are accessed by index
# * texts are accessed by index
# * non-matched texts are kept (to preserve index order)
listSearch = AhoCorasickSearchList(keywords, list(c("What in", "the world"), c("is"), "secret about", "the Magic Word?"))
stopifnot(listEquals(listSearch[[1]][[1]], list()))
stopifnot(listEquals(listSearch[[1]][[2]][[1]], list(keyword="the", offset=1)))
stopifnot(listEquals(listSearch[[2]][[1]][[1]], list(keyword="is", offset=1)))
stopifnot(listEquals(listSearch[[3]], list()))
stopifnot(listEquals(listSearch[[4]][[1]][[1]], list(keyword="the", offset=1)))
stopifnot(listEquals(listSearch[[4]][[1]][[2]], list(keyword="Magic", offset=5)))
stopifnot(listEquals(listSearch[[4]][[1]][[3]], list(keyword="Word", offset=11)))
# named search of a list of lists
# * sublists are accessed by name
# * matched texts are accessed by name
# * non-matched texts are dropped
namedSearch = AhoCorasickSearchList(keywords, list(subject=c(phrase1="What in", phrase2="the world"),
verb=c(phrase1="is"),
predicate1=c(phrase1="secret about"),
predicate2=c(phrase1="the Magic Word?")))
stopifnot(listEquals(namedSearch$subject$phrase2[[1]], list(keyword="the", offset=1)))
stopifnot(listEquals(namedSearch$verb$phrase1[[1]], list(keyword="is", offset=1)))
stopifnot(listEquals(namedSearch$predicate1, list()))
stopifnot(listEquals(namedSearch$predicate2$phrase1[[1]], list(keyword="the", offset=1)))
stopifnot(listEquals(namedSearch$predicate2$phrase1[[2]], list(keyword="Magic", offset=5)))
stopifnot(listEquals(namedSearch$predicate2$phrase1[[3]], list(keyword="Word", offset=11)))
# named search of multiple texts in a single list with keyword grouping and aminoacid alphabet
# * all matches to a keyword are accessed by name
# * non-matched keywords are dropped
proteins = c(protein1="PEPTIDEPEPTIDEDADADARARARARAKEKEKEKEPEPTIDE",
protein2="DERPADERPAPEWPEWPEEPEERAWRAWWARRAGTAGPEPTIDEKESEQUENCE")
peptides = c("PEPTIDE", "DERPA", "SEQUENCE", "KEKE", "PEPPIE")
peptideSearch = AhoCorasickSearch(peptides, proteins, alphabet="aminoacid", groupByKeyword=T)
stopifnot(listEquals(peptideSearch$PEPTIDE, list(list(keyword="protein1", offset=1),
list(keyword="protein1", offset=8),
list(keyword="protein1", offset=37),
list(keyword="protein2", offset=38))))
stopifnot(listEquals(peptideSearch$DERPA, list(list(keyword="protein2", offset=1),
list(keyword="protein2", offset=6))))
stopifnot(listEquals(peptideSearch$SEQUENCE, list(list(keyword="protein2", offset=47))))
stopifnot(listEquals(peptideSearch$KEKE, list(list(keyword="protein1", offset=29),
list(keyword="protein1", offset=31),
list(keyword="protein1", offset=33))))
stopifnot(listEquals(peptideSearch$PEPPIE, NULL))
# grouping by keyword without text names: offsets are given without reference to the text
names(proteins) = NULL
peptideSearch = AhoCorasickSearch(peptides, proteins, groupByKeyword=T)
stopifnot(listEquals(peptideSearch$PEPTIDE, list(1, 8, 37, 38)))
stopifnot(listEquals(peptideSearch$DERPA, list(1, 6)))
stopifnot(listEquals(peptideSearch$SEQUENCE, list(47)))
stopifnot(listEquals(peptideSearch$KEKE, list(29, 31, 33)))
I have a R code that I am trying to run in a server. But it is stopping in the middle/get frozen probably because of memory limitation. The data files are huge/massive (one has 20 million lines) and if you look at the double for loop in the code, length(ratSplit) = 281 and length(humanSplit) = 36. The data has specific data of human and rats' genes and human has 36 replicates, while rat has 281. So, the loop is basically 281*36 steps. What I want to do is to process data using the function getGeneType and see how different/independent are the expression of different replicate combinations. Using Fisher's test. The data rat_processed_7_25_FDR_05.out looks like this :
2 Sptbn1 114201107 114200202 chr14|Sptbn1:114201107|Sptbn1:114200202|reg|- 2 Thymus_M_GSM1328751 reg
2 Ndufb7 35680273 35683909 chr19|Ndufb7:35680273|Ndufb7:35683909|reg|+ 2 Thymus_M_GSM1328751 rev
2 Ndufb10 13906408 13906289 chr10|Ndufb10:13906408|Ndufb10:13906289|reg|- 2 Thymus_M_GSM1328751 reg
3 Cdc14b 1719665 1719190 chr17|Cdc14b:1719665|Cdc14b:1719190|reg|- 3 Thymus_M_GSM1328751 reg
and the data fetal_output_7_2.out has the form
SPTLC2 78018438 77987924 chr14|SPTLC2:78018438|SPTLC2:77987924|reg|- 11 Fetal_Brain_408_AGTCAA_L006_R1_report.txt reg
EXOSC1 99202993 99201016 chr10|EXOSC1:99202993|EXOSC1:99201016|rev|- 5 Fetal_Brain_408_AGTCAA_L006_R1_report.txt reg
SHMT2 57627893 57628016 chr12|SHMT2:57627893|SHMT2:57628016|reg|+ 8 Fetal_Brain_408_AGTCAA_L006_R1_report.txt reg
ZNF510 99538281 99537128 chr9|ZNF510:99538281|ZNF510:99537128|reg|- 8 Fetal_Brain_408_AGTCAA_L006_R1_report.txt reg
PPFIBP1 27820253 27824363 chr12|PPFIBP1:27820253|PPFIBP1:27824363|reg|+ 10 Fetal_Brain_408_AGTCAA_L006_R1_report.txt reg
Now I have few questions on how to make this more efficient. I think when I run this code, R takes up lots of memory that ultimately causes problems. I am wondering if there is any way of doing this more efficiently
Another possibility is the usage of double for-loop'. Will sapply help? In that case, how should I apply sapply?
At the end I want to convert result into a csv file. I know this is a bit overwhelming to put code like this. But any optimization/efficient coding/programming will be A LOT! I really need to run the whole thing at least one to get the data soon.
#this one compares reg vs rev
date()
ratRawData <- read.table("rat_processed_7_25_FDR_05.out",col.names = c("alignment", "ratGene", "start", "end", "chrom", "align", "ratReplicate", "RNAtype"), fill = TRUE)
humanRawData <- read.table("fetal_output_7_2.out", col.names = c("humanGene", "start", "end", "chrom", "alignment", "humanReplicate", "RNAtype"), fill = TRUE)
geneList <- read.table("geneList.txt", col.names = c("human", "rat"), sep = ',')
#keeping only information about gene, alignment number, replicate and RNAtype, discard other columns
ratRawData <- ratRawData[,c("ratGene", "ratReplicate", "alignment", "RNAtype")]
humanRawData <- humanRawData[, c( "humanGene", "humanReplicate", "alignment", "RNAtype")]
#function to capitalize
capitalize <- function(x){
capital <- toupper(x) ## capitalize
paste0(capital)
}
#capitalizing the rna type naming for rat. So, reg ->REG, dup ->DUP, rev ->REV
#doing this to make data manipulation for making contingency table easier.
levels(ratRawData$RNAtype) <- capitalize(levels(ratRawData$RNAtype))
#spliting data in replicates
ratSplit <- split(ratRawData, ratRawData$ratReplicate)
humanSplit <- split(humanRawData, humanRawData$humanReplicate)
print("done splitting")
#HyRy :when some gene has only reg, rev , REG, REV
#HnRy : when some gene has only reg,REG,REV
#HyRn : add 1 when some gene has only reg,rev,REG
#HnRn : add 1 when some gene has only reg,REG
#function to be used to aggregate
getGeneType <- function(types) {
types <- as.character(types)
if ('rev' %in% types) {
return(ifelse(('REV' %in% types), 'HyRy', 'HyRn'))
}
else {
return(ifelse(('REV' %in% types), 'HnRy', 'HnRn'))
}
}
#logical function to see whether x is integer(0) ..It's used the for loop bellow in case any one HmYn is equal to zero
is.integer0 <- function(x) {
is.integer(x) && length(x) == 0L
}
result <- data.frame(humanReplicate = "human_replicate", ratReplicate = "rat_replicate", pvalue = "p-value", alternative = "alternative_hypothesis",
Conf.int1 = "conf.int1", Conf.int2 ="conf.int2", oddratio = "Odd_Ratio")
for(i in 1:length(ratSplit)) {
for(j in 1:length(humanSplit)) {
ratReplicateName <- names(ratSplit[i])
humanReplicateName <- names(humanSplit[j])
#merging above two based on the one-to-one gene mapping as in geneList defined above.
mergedHumanData <-merge(geneList,humanSplit[[j]], by.x = "human", by.y = "humanGene")
mergedRatData <- merge(geneList, ratSplit[[i]], by.x = "rat", by.y = "ratGene")
mergedHumanData <- mergedHumanData[,c(1,2,4,5)] #rearrange column
mergedRatData <- mergedRatData[,c(2,1,4,5)] #rearrange column
mergedHumanRatData <- rbind(mergedHumanData,mergedRatData) #now the columns are "human", "rat", "alignment", "RNAtype"
agg <- aggregate(RNAtype ~ human+rat, data= mergedHumanRatData, FUN=getGeneType) #agg to make HmYn form
HmRnTable <- table(agg$RNAtype) #table of HmRn ie RNAtype in human and rat.
#now assign these numbers to variables HmYn. Consider cases when some form of HmRy is not present in the table. That's why
#is.integer0 function is used
HyRy <- ifelse(is.integer0(HmRnTable[names(HmRnTable) == "HyRy"]), 0, HmRnTable[names(HmRnTable) == "HyRy"][[1]])
HnRn <- ifelse(is.integer0(HmRnTable[names(HmRnTable) == "HnRn"]), 0, HmRnTable[names(HmRnTable) == "HnRn"][[1]])
HyRn <- ifelse(is.integer0(HmRnTable[names(HmRnTable) == "HyRn"]), 0, HmRnTable[names(HmRnTable) == "HyRn"][[1]])
HnRy <- ifelse(is.integer0(HmRnTable[names(HmRnTable) == "HnRy"]), 0, HmRnTable[names(HmRnTable) == "HnRy"][[1]])
contingencyTable <- matrix(c(HnRn,HnRy,HyRn,HyRy), nrow = 2)
# contingencyTable:
# HnRn --|--HyRn
# |------|-----|
# HnRy --|-- HyRy
#
fisherTest <- fisher.test(contingencyTable)
#make new line out of the result of fisherTest
newLine <- data.frame(t(c(humanReplicate = humanReplicateName, ratReplicate = ratReplicateName, pvalue = fisherTest$p,
alternative = fisherTest$alternative, Conf.int1 = fisherTest$conf.int[1], Conf.int2 =fisherTest$conf.int[2],
oddratio = fisherTest$estimate[[1]])))
result <-rbind(result,newLine) #append newline to result
if(j%%10 = 0) print(c(i,j))
}
}
write.table(result, file = "compareRegAndRev.csv", row.names = FALSE, append = FALSE, col.names = TRUE, sep = ",")
Referring to the accepted answer to Monitor memory usage in R, the amount of memory used by R can be tracked with gc().
If the script is, indeed, running short of memory (which would not surprise me), the easiest way to resolve the problem would be to move the write.table() from the outside to the inside of the loop, to replace the rbind(). It would just be necessary to create a new file name for the CSV file that is written from each output, e.g. by:
csvFileName <- sprintf("compareRegAndRev%03d_%03d.csv",i,j)
If the CSV files are written without headers, they could then be concatenated separately outside R (e.g. using cat in Unix) and the header added later.
While this approach might succeed in creating the CSV file that is sought, it is possible that file might be too big to process subsequently. If so, it may be preferable to process the CSV files individually, rather than concatenating them at all.
As a fairly new R programmer I seem to have run into a strange problem - probably my inexperience with R
After reading and merging successive files into a single data frame, I find that order does not sort the data as expected.
I have multiple references in each file but each file refers to measurement data obtained at a different time.
Here's the code
library(reshape)
# Enter file name to Read & Save data
FileName=readline("Enter File name:\n")
# Find first occurance of file
for ( round1 in 1 : 6) {
ReadFile=paste(round1,"C_",FileName,"_Stats.csv", sep="")
if (file.exists(ReadFile))
break
}
x = data.frame(read.csv(ReadFile, header=TRUE),rnd=round1)
for ( round2 in (round1+1) : 6) {
#
ReadFile=paste(round2,"C_",FileName,"_Stats.csv", sep="")
if (file.exists(ReadFile)) {
y = data.frame(read.csv(ReadFile, header=TRUE),rnd = round2)
if (round2 == (round1 +1))
z=data.frame(merge(x,y,all=TRUE))
z=data.frame(merge(y,z,all=TRUE))
}
}
ordered = order(z$lab_id)
results = z[ordered,]
res = data.frame( lab=results[,"lab_id"],bw=results[,"ZBW"],wi=results[,"ZWI"],pf_zbw=0,pf_zwi=0,r = results[,"rnd"])
#
# Establish no of samples recorded
nsmpls = length(res[,c("lab")])
# Evaluate Z_scores for Between Lab Results
for ( i in 1 : nsmpls) {
if (res[i,"bw"] > 3 | res[i,"bw"] < -3)
res[i,"pf_zbw"]=1
}
# Evaluate Z_scores for Within Lab Results
for ( i in 1 : nsmpls) {
if (res[i,"wi"] > 3 | res[i,"wi"] < -3)
res[i,"pf_zwi"]=1
}
dd = melt(res, id=c("lab","r"), "pf_zbw")
b = cast(dd, lab ~ r)
If anyone could see why the ordering only works for about 55 of 70 records and could steer me in the right direction I would be obliged
Thanks very much
Check whether z$lab_id is a factor (with is.factor(z$lab_id)).
If it is, try
z$lab_id <- as.character(z$lab_id)
if it is supposed to be a character vector; or
z$lab_id <- as.numeric(as.character(z$lab_id))
if it is supposed to be a numeric vector.
Then order it again.
Ps. I had previously put these in the comments.