How To Select Only Human Mirnas From Affy Chip While Data Analysis Using R

Question

[Shafi]

I am new to R and want to analyze miRNA data set of 3 groups. I am able to normalize, designing matrix and contrast matrix and differential expression. But what I am facing is affy chip has lot of other miRNA probes which I guess I am not able to filter them out. I just want to select human miRNAs from chip and do analysis and compare the three groups. Can any one help me in this

[Qdjm]

Shari, it would help to mention which Affy chip you are using. If they are annotated using miRBase annotations, then the human miRNA names should all start with "hsa"

[Shafi]

its mirrna20 from affymetrix

[Shafi]

topTable(fit2, coef=1, adjust="fdr", sort.by="B", number=20) ID logFC AveExpr t P.Value adj.P.Val B 20398 zma-miR156e_st 0.3198252 0.4998068 4.487144 3.782516e-05 0.6469183 0.62186254 18047 pvu-miR482-star_st -0.2559547 0.5195567 -4.233259 8.925328e-05 0.6469183 0.09110598 2995 ame-miR-316_st -0.3233149 0.4537794 -4.029978 1.748058e-04 0.6469183 -0.32656880 12019 hsa-miR-3119_st -0.3040332 0.4469291 -4.016812 1.824927e-04 0.6469183 -0.35336278 2088 ENSG00000252228_x_st -0.3748713 0.6650043 -3.990783 1.986619e-04 0.6

[Shafi]

you can not only has but others are dominant

[Shafi]

you can see not only has but other species are dominant in topTAble

[Shafi]

you can see not only has but other species are dominant in topTAble

[Shafi]

ExpressionSet (storageMode: lockedEnvironment) assayData: 20706 features, 55 samples element names: exprs protocolData sampleNames: EC1_(miRNA).CEL EC21_(miRNA).CEL ... YC85_(miRNA).CEL (55 total) varLabels: ScanDate varMetadata: labelDescription phenoData sampleNames: EC1_(miRNA).CEL EC21_(miRNA).CEL ... YC85_(miRNA).CEL (55 total) varLabels: sample varMetadata: labelDescription featureData: none experimentData: use 'experimentData(object)' Annotation: mirna20

[Qdjm]

Shafi, looks like you have "hsa" miRNAs. So filter out all miRNAs whose names don't start with hsa. Then count the number of miRNAs left -- if it's less than 300, you probably overfiltered and if its more than about 1500, you haven't filtered enough.

[Shafi]

yes that is what I want to do. Can you help me how I can select only has miRNAs? I appreciate you kind help

[Shafi]

yes that is what I want to do. Can you help me how I can select only has miRNAs? I appreciate you kind help

[Shafi]

yes that is what I want to do. Can you help me how I can select only "hsa" miRNAs? I appreciate you kind help

Answer 1

[Agatha]

If I understood correctly, you want to filter them out-> just the human miRNAs. You can use grep for that Here is an example (loads a fasta file containing the hairpins annotated in miRBase and filters the human ones)

library("Biostrings")
aux <- read.RNAStringSet("hairpin.fa")
names(aux)
grep("hsa",names(aux))
aux[grep("hsa",names(aux))]

You can adapt it for your task. Hope this helps.

[Agatha]

@Shafi- you cannot open the file since it is not saved locally on your computer. I do not have access to your code so I gave you an example on how to extract rows from a data object containing a specific substring, in this case, "hsa" the default start string for the miRNAs annotated in mirBase

[Shafi]

Error in .Call2("new_input_ExternalFilePtr", fp, PACKAGE = "Biostrings") : cannot open file 'hairpin.fa'

[Shafi]

Did not work! What should I do....

Answer 2

[Shafi]

Hi

I just want to study differential expression of human miRNAs and compare 3 gruops. I have used affy chips (mirna20). Just 2 weeks before I started learning R. Dont know much. During these days I learned how to load, read, normalize(RMA) data. I will appreciate if some one will help me out or direct me towards relevant example. Thanks in advance

Answer 3

[User 2925]

Hello Agatha

would the same example code work with biopython? i have also a hairpin.fa file from miRbase and am only interested in the hsa-mir-*** miRNAs. how can i extract them with the id,and sequences?

Answer 4

[Shafi]

Hi

I am also new to R. I used following script to grep only human miRNAs.

# grep("hsa", featureNames(eset))->hsa

# eset.hsa<-eset[hsa,]

Hope it will work for you