T-Test Filtering Problem In R (Simpleaffy)

Question

[Andrii]

Dear all! I need your help!!! Situation is next: I tried to perform t-test in R. I applied function

result.0vs4<-pairwise.comparison(eset,"time",c("control","4 h"))

to the expression set object (Affymetrix)

result.0vs4
........
AFFX-r2-P1-cre-5_at              AFFX-ThrX-3_at 
           1.571096e-01                2.742916e-01 
         AFFX-ThrX-5_at              AFFX-ThrX-M_at 
           2.940117e-01                8.981617e-01 
        AFFX-TrpnX-3_at             AFFX-TrpnX-5_at 
           6.314353e-01                4.371810e-01 
        AFFX-TrpnX-M_at 
           4.755571e-01 
attr(,"Csingle")
[1] TRUE

Slot "calls":

<0 x 0 matrix>

Slot "group":

[1] "time"

Slot "members":

[1] "control" "4 h"

Slot "pData":

          sample     agent    time
GSM442026 GSM442026 untreated control
GSM442027 GSM442027 untreated control
GSM442028 GSM442028 untreated control
GSM442037 GSM442037  TGF-beta     4 h
GSM442038 GSM442038  TGF-beta     4 h
GSM442039 GSM442039  TGF-beta     4 h

...............................

After that everything was fine, every slot of PairComp object (resulted object after pairwise.comparison function application) had information about fold change, means etc. Till this point no problem, but after applying filtering each slot was empty:-( I used filter contritions as follows: p> 0.001, fold change>2.

result.0vs4.f=pairwise.filter(result.0vs4,tt=0.001,fc=2)
result.0vs4.f
An object of class "PairComp"

Slot "means":

control 4 h

Slot "fc":

named numeric(0)

Slot "tt":

named numeric(0)

Slot "calls":

<0 x 0 matrix>

Slot "group":

[1] "time"

Slot "members":

[1] "control" "4 h"

Slot "pData":

          sample     agent    time
GSM442026 GSM442026 untreated control
GSM442027 GSM442027 untreated control
GSM442028 GSM442028 untreated control
GSM442037 GSM442037  TGF-beta     4 h
GSM442038 GSM442038  TGF-beta     4 h
GSM442039 GSM442039  TGF-beta     4 h

..................

I set less strict parameters and everything was the same:-( If somebody knows the reason of that don't stand aside :-) Thank you in advance! Andrii

Answer 1

[Neilfws]

I think you mean p < 0.001, not p > 0.001. Anyway, the most likely reason for the empty slots - assuming the code ran without errors - is simply that no probeset matched your filter conditions.