maSigPro warnings in glm.fit

Question

[Assa Yeroslaviz]

Hi,

I am working on a RNA-Seq data set from drosophila melanogaster time course analysis. I am trying to computer a regression fit for each of the genes.

When computing the p-values I am getting multiple warnings like this:

There were 50 or more warnings (use warnings() to see the first 50)
> warnings()
Warning messages:
1: glm.fit: algorithm did not converge
2: glm.fit: algorithm did not converge

I was wondering what causes this warning and is it essential to change some parameters for better results.

Thanks
Assa

My edesign is this:

sampleName    Time    Replicates    Group
IFM_Myoblast_1    0    1    1
IFM_Myoblast_2    0    1    1
IFM_Myoblast_3    0    1    1
IFM16h_1    16    2    1
IFM16h_2    16    2    1
IFM16h_3    16    2    1
IFM24h_1    24    3    1
IFM24h_2    24    3    1
IFM24h_3    24    3    1
IFM30h_1    30    4    1
IFM30h_2    30    4    1
IFM48h_1    48    5    1
IFM48h_2    48    5    1
IFM48h_3    48    5    1
IFM72h_1    72    6    1
IFM72h_2    72    6    1
IFM90h_1    90    7    1
IFM90h_2    90    7    1
IFM100h_1    100    8    1
IFM100h_2    100    8    1

[Stephanie]

Hi Assa,

Did you get any answer to your question? I have the same kind of design, also in D. melanogaster, and get the same error message.

Also, when I use next the T.fit() function, I get the following error:

[1] "fitting gene 100 out of 1061"
[1] "fitting gene 200 out of 1061"
[1] "fitting gene 300 out of 1061"
[1] "fitting gene 400 out of 1061"
[1] "fitting gene 500 out of 1061"
[1] "fitting gene 600 out of 1061"
[1] "fitting gene 700 out of 1061"
[1] "fitting gene 800 out of 1061"
Error: object 'lastname' not found

I don't know if it is due to the warning message or if there is another problem.

Thank your for your help.

Stephanie

[Assa Yeroslaviz]

sorry, but no. As it was only a warning, I sort of ignored it.