Learn how easy it is to sync an existing GitHub or Google Code repo to a SourceForge project! See Demo
Dear TMEV team!
Downloaded, installed and tested TMEV_4_6_2_r2536_win for SAM-R and LIMMA-R.
Test file is .tdms, 4 samples, 3-4 replicates each, 9553 genes, normalized log2 ratios, no missing values.
SAM-R settings: use R; Number of permutations=calculated suggested; Select S0=using Tusher et al. method; Calculate q-values?=Yes.
One-class test (one sample with 4 replicates selected from the list):
In both Significant Genes and Non-significant Genes Tables:
column Numerator (r)=0.0 for all genes
column Denominator (s+s0)=0.0 for all genes
column q-value (%)=0.0 for all genes
First two columns, apparently, should have some meaningful values, and as genes were arranged into Significant Genes and Non-significant Genes, there must be some meaningful q-values too.
Two-class unpaired test (two samples with 4 replicates each selected from the list) – exactly the same problems with the same columns.
Also, for both tests, in General Information-Computed Quantities, Computed Exchangeability Factor s0=0.0, s0 Percentile=0.0, Pi0Hat=0.0, which is unlikely.
Is there a problem with reading the output data from samr to the tables? Interestingly, when I run exactly the same analysis, but NOT using the R, all output columns in all gene tables has meaningful values, as well as in General Information-Computed Quantities.
LIMMA-R settings : Significance Level: Alpha=.05
One Class test (one sample with 4 replicates selected from the list):
Gene Tables output has no columns for significance values, adj-p-values, log fold change, t-statistic, log-odds, but has irrelevant column F-values, containing only 0.0 (applicable to 3 or more samples).
Two Class test ((two samples with 4 replicates each selected from the list)-all columns are right, but again has this irrelevant column F-values, containing only 0.0 (applicable to 3 and more samples).
Is there a problem with reading the output data from limma-r to the tables for one-class experiment? And why the column F-value is popping up where it shouldn't?
For me it looks like these two very useful additions to the data analysis process are again not working properly, the same as I reported for 4_6_1. Or is it only me who's having these problems?
Hi, We will definitely look into this and get back to you.