[BioC] FW: duplicates, technical and biological replicates + dividing a microarray into two parts
Staninska, Ana, Dr.
ana.staninska at helmholtz-muenchen.de
Mon Feb 1 16:50:08 CET 2010
Dear BioConductor team,
I am working on a statistical analysis of 2 color Genepix data. For this analysis, I am using the Limma package.
I am trying to find expressed genes in the treated samples vs the non-treated samples, where each sample is a leaf from a particular tree.
I have 5 biological replicates (5 treated and 5 untreated), and for each biological replicates- -- 2 technical replicates, which are dye swapped arrays (10 microarrays total).
The spots on each microarray are duplicated: on the left side and on the right side
So...The experiment design involves 3 kinds of replicates: within array replicate (duplicate spots), technical replicates (as dye swap), and biological replicates.
I know that at the moment it is not possible with limma to treat this kind of experiment, but I have an idea how to avoid duplicate spots (within array replicates), if that is possible.
Here I need your help.
There are 8x4 (8 rows, 4 columns) print tip groups on the microarrays, and each print tip group is of size 12x8 (but i think that is not relevant for now).
The experiment is designed such that the left hand side of the microarray and the right hand side are identical. Basically the duplicate spots are spotted on the left and on the right hand side of the array (if the blocks are numbered 1 through 32, then 1 and 3 are same, 2 and 4, 5 and 7, 6 and 8 etc….) .
So if somehow I can divide my microarray into two peaces and treat the peaces as two separate microarrays, then I will be able to avoid the duplicate spots, and only deal with technical and biological replicates.
So if my original microarray consists of 32 blocs (print tip groups), I would like the two new microarrays, called Left_microarray and Right_microarray each to contain 16 blocks, such that the blocks
1,2,5,6,9,10,13,14,17,18,21,22,25,26,29,30 to be in Left_microarray and the remaining blocks 3,4,7,8,11,12,15,16,19,20,23,24,27,28,31,32 to be in the Right microarray.
Is this possible?
If it is, could you please help me and tell me how to do this?
Just in case, I am also sending my R code for the experiment.
Thank you very much in advance
Ana Staninska
Institute of Biomathematics and Biometry
Helmholtz-Zentrum München
München, Deutschand
The R-code of the experiment:
I tried all the possible cases to deal with the experiment: averaging the within array replicates, treating biological as technical replicates, or treating technIcal as biological replicates.
After I ran the R code, I compared the results with the qRT-PCR results previously done for the experiments. The comparison was done such that I took the sum of the absolute values of the subtraction of log FC form qRT-PCR and logFC from my analysis.
It turned out that treating technical as biological replicates was the worst possibility, but treating biological as technical replicates was the best.
> targets <- readTargets("Lysi_270706.txt")
>
> myfun<-function(x) {
+ nored<-abs(x[,"F635 Median"] + x[,"F635 Mean"]) !=0
+ nogreen<-abs(x[, "F532 Median"]+x[,"F532 Mean"]) !=0
+ as.numeric(nogreen & nored)
+ }
>
> RGa <- read.maimages(targets, source="genepix", wt.fun=myfun, other.columns=c("F635 SD","B635 SD","F532 SD","B532 SD","B532 Mean","B635 Mean","F Pixels","B Pixels"))
Read Met270706_1_60308.gpr
Read Met270706_dw1_110308.gpr
Read Met270706_2_060308.gpr
Read Met270706_dw2_110308.gpr
Read Met270706_3_060308.gpr
Read Met270706_dw3_120308.gpr
Read Met270706_4_060308.gpr
Read Met270706_dw4_120308.gpr
Read Met270706_5_060308.gpr
Read Met270706_dw5_120308.gpr
Read Met270706_6_060308.gpr
Read Met270706_dw6_120308.gpr
Read Met270706_7_110308.gpr
Read Met270706_dw7_120308.gpr
Read Met270706_8_220408.gpr
Read Met270706_dw8_120308.gpr
Read Met270706_9_110308.gpr
Read Met270706_dw9_120308.gpr
Read Met270706_10_110308.gpr
Read Met270706_dw10_120308.gpr
>
> RG.ne10b <-backgroundCorrect(RGa, method="normexp", , normexp.method="mle", offset=10)
Green channel
Corrected array 1
Corrected array 2
Corrected array 3
Corrected array 4
Corrected array 5
Corrected array 6
Corrected array 7
Corrected array 8
Corrected array 9
Corrected array 10
Corrected array 11
Corrected array 12
Corrected array 13
Corrected array 14
Corrected array 15
Corrected array 16
Corrected array 17
Corrected array 18
Corrected array 19
Corrected array 20
Red channel
Corrected array 1
Corrected array 2
Corrected array 3
Corrected array 4
Corrected array 5
Corrected array 6
Corrected array 7
Corrected array 8
Corrected array 9
Corrected array 10
Corrected array 11
Corrected array 12
Corrected array 13
Corrected array 14
Corrected array 15
Corrected array 16
Corrected array 17
Corrected array 18
Corrected array 19
Corrected array 20
>
> MA_l.ne10b <- normalizeWithinArrays(RG.ne10b, method="loess")
>
> #################################################################
> ### Average of the Duplicate Spots ###
> #################################################################
>
> MAa_l.ne10b <- avedups(MA_l.ne10b, ndups=2, spacing=192)
> design <- modelMatrix(targets, ref="wt")
Found unique target names:
mu wt
> biolrep<-c(1,1,2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10)
>
> corfita_l.ne10b<-duplicateCorrelation(MAa_l.ne10b, design, block=biolrep)
>
> fita_l.ne10b<-lmFit(MAa_l.ne10b, design, block=biolrep, cor=corfita_l.ne10b$consensus)
>
> fita_l.ne10b<-eBayes(fita_l.ne10b)
>
> TTa_l.ne10b<-topTable(fita_l.ne10b,coef=1, number=1600, adjust="BH")
> write.csv(TTa_l.ne10b, file="BC_Lysi_270706a_TTa_l_ne10b.csv")
>
> ################################################################
> ### BIOLOGICAL AS TECHNICAL ######
> ################################################################
>
> corfit_l.ne10b<-duplicateCorrelation(MA_l.ne10b, ndups=2, spacing=192)
>
> fitbt_l.ne10b<-lmFit(MA_l.ne10b, design, ndups=2, spacing=192, cor=corfit_l.ne10b$consensus)
>
> fitbt_l.ne10b<-eBayes(fitbt_l.ne10b)
>
> TTbt_l.ne10b<-topTable(fitbt_l.ne10b,coef=1, number=1600, adjust="BH")
> write.csv(TTbt_l.ne10b, file="BC_Lysi_270706a_TTbt_l_ne10b.csv")
>
> ###############################################################
> #### TECNICAL AS BIOLOGICAL ####
> ###############################################################
>
>
>
> design1<-cbind(
+ nt1=c( 1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0),
+ tr1=c(0, -1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0),
+ nt2=c(0,0, 1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0),
+ tr2=c(0,0,0, -1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0),
+ nt3=c(0,0,0,0, 1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0),
+ tr3=c(0,0,0,0,0, -1,0,0,0,0,0,0,0,0,0,0,0,0,0,0),
+ nt4=c(0,0,0,0,0,0, 1,0,0,0,0,0,0,0,0,0,0,0,0,0),
+ tr4=c(0,0,0,0,0,0,0, -1,0,0,0,0,0,0,0,0,0,0,0,0),
+ nt5=c(0,0,0,0,0,0,0,0, 1,0,0,0,0,0,0,0,0,0,0,0),
+ tr5=c(0,0,0,0,0,0,0,0,0, -1,0,0,0,0,0,0,0,0,0,0),
+ nt6=c(0,0,0,0,0,0,0,0,0,0, 1,0,0,0,0,0,0,0,0,0),
+ tr6=c(0,0,0,0,0,0,0,0,0,0,0, -1,0,0,0,0,0,0,0,0),
+ nt7=c(0,0,0,0,0,0,0,0,0,0,0,0, 1,0,0,0,0,0,0,0),
+ tr7=c(0,0,0,0,0,0,0,0,0,0,0,0,0, -1,0,0,0,0,0,0),
+ nt8=c(0,0,0,0,0,0,0,0,0,0,0,0,0,0, 1,0,0,0,0,0),
+ tr8=c(0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, -1,0,0,0,0),
+ nt9=c(0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 1,0,0,0),
+ tr9=c(0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, -1,0,0),
+ nt10=c(0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 1,0),
+ tr10=c(0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, -1))
>
> fittb_l.ne10b<-lmFit(MA_l.ne10b, design1, ndups=2, spacing=192,cor=corfit_l.ne10b$consensus)
Warning message:
Partial NA coefficients for 160 probe(s)
>
> fittb_l.ne10b<-eBayes(fittb_l.ne10b)
>
> TTtb_l.ne10b<-topTable(fittb_l.ne10b,coef=1, number=1600, adjust="BH")
> write.csv(TTtb_l.ne10b, file="BC_Lysi_270706a_TTtb_l_ne10b.csv")
>
Ana Staninska
Institute of Biomathematics and Biometry
Helmholtz-Zentrum München
München, Deutschand
More information about the Bioconductor
mailing list