## User: zoppoli pietro

Reputation:
10
Status:
New User
Location:
United States
Last seen:
1 week ago
Joined:
7 years, 4 months ago
Email:
z*************@gmail.com

#### Posts by zoppoli pietro

<prev • 15 results • page 1 of 2 • next >
1
66
views
1
... Dear Aaron, thanks for your answer. I tried your solution but with no luck. Maybe it's related to the used of the offset from EDAseq package together with the RUVg function from RUVSeq package. As you can read in the previous post my design is  design <- model.matrix(~conditions + W_1, data= ...
written 17 days ago by zoppoli pietro10
1
66
views
1
... Hi Aaron, thanks a lot for your answer, you made me reflect on the many reasons the values can be different. I expected a minor difference in FC as you state in your answer BUT I find  FCs with different sign comparing the results of my analysis and the cpm. Let me explain that i'm using both EDA ...
written 18 days ago by zoppoli pietro10
1
66
views
1
... I have a DGEGLM object from edgeR called fit2 fit2 <- glmFit(counts(Data), design, disp2$tagwise.dispersion,offset=-offst(dataOffset)) I want the logFC values so logFC <- fit2$coefficients[geneA,2] # first column intercept, second column is my comparison fit2$fitted.values[geneA,] # i ... written 19 days ago by zoppoli pietro10 • updated 18 days ago by Aaron Lun21k 1 answers 169 views 1 answers ... Hi Davide, if I understand correctly, in my example:$Y is the fit2$counts$W$is the lrt2$coefficients (Ngenes X 3conditions) $\alpha$ is the design$W_1 (W_1 is a vector result of RUVg) (Nsamples X 1) so MatrixByDavide <- fit_TvsN$coefficients[,W_1]%*%t(design$W_1) MatrixToMakePlots &l ... written 20 days ago by zoppoli pietro10 1 answers 169 views 1 answers ... Hi Davide, my bad but i really can't understand the formula$Y - W alpha$. Can you please explain it to me ? Please consider that in my example fit2 is the result of the glmFit having pData(set1)$ W_1 integrated into design and set1 is the result of RUVg. Sorry to bother but i'd really like to fu ...
written 6 weeks ago by zoppoli pietro10
1
169
views
1
... Hi Davide, thanks for your reply. I explained myself the discrepancy in the same way BUT this means I can't use boxplot or similar plots to show the data. I can do a HM of the genes using the edgeR "logFC" but i have no way to show the distribution of the samples around that value, right ? By the ...
written 6 weeks ago by zoppoli pietro10
1
169
views
1
... Hi Davide, I'm going to visualize the results of the analysis using a simple boxplot and I find a myself troubled to decide which matrix to use. Let's say I get lrt2 from glmLRT and I want to plot geneA   contrast_BvsA <- makeContrasts(TvsN=conditionsB-conditionA, levels=design) contrast_Bv ...
written 6 weeks ago by zoppoli pietro10
1
169
views
1
... Hi Davide, following your tips I changed this dispersion: disp <- estimateGLMCommonDisp(counts(dataOffset), design, offset=-offst(dataOffset)) with this dispersion disp2 <- estimateDisp(counts(dataOffset),design, offset=-offst(dataOffset)) and changed the fit step: fit <- glmFit(c ...
written 10 weeks ago by zoppoli pietro10
1
169
views
1
... Hi, I'd like to use together EDASeq,RUVSeq and edgeR I'd like to know if It's correct to run the folling code According to EDASeq manual: dataOffset <- withinLaneNormalization(data,"gc",which="upper",offset=TRUE) dataOffset <- betweenLaneNormalization(dataOffset,which="upper",offset=TRUE ...
written 10 weeks ago by zoppoli pietro10 • updated 10 weeks ago by davide risso780
2
274
views
2
... Sorry to bother but I'd like to know if it's correct to pass the normCounts(dataOffset) obtained in the previous example directly to voom in order to to process RNA-Seq data prior to linear modelling in limma. I'm trying to make a comparison on the different pipelines to make a DEA from my RNAseq d ...
written 7 months ago by zoppoli pietro10

#### Latest awards to zoppoli pietro

No awards yet. Soon to come :-)

Content
Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.