Hello fellow researchers,

using masigpro i encountered an error which I cant seem to fix. Hopefully someone is working with masigpro as well or the developers are still around.

As you can see the p.vector function does what it should with "ncounts" beeing the normalized counts and d_1 the design matrix.

Calling T.fit() however gives me the following error:

> NBp_1 <- p.vector(ncounts, d_1, counts=TRUE,min.obs=3)

[1] "fitting gene 100 out of 165"
> NBt_1 <- T.fit(NBp_1)
[1] "fitting gene 100 out of 120"
Error in terms.formula(formula, data = data) : 
  '.' in formula and no 'data' argument

To show you the data used as input:

> d_1

$dis
           group.2vsgroup.1 time timexgroup.2 time2 time2xgroup.2
2h                        1    2            2     4             4
4h                        1    4            4    16            16
8h                        1    8            8    64            64
10h                       1   10           10   100           100
12h                       1   12           12   144           144
14h                       1   14           14   196           196
24h                       1   24           24   576           576
h2Control                 0    2            0     4             0
h4Control                 0    4            0    16             0
h8Control                 0    8            0    64             0
h10Control                0   10            0   100             0
h12Control                0   12            0   144             0
h14Control                0   14            0   196             0
h24Control                0   24            0   576             0

$groups.vector
[1] "group.2vsgroup.1" "group.1"          "group.2vsgroup.1" "group.1"          "group.2vsgroup.1"

$edesign
           time replicates group.1 group.2
2h            2          1       0       1
4h            4          2       0       1
8h            8          3       0       1
10h          10          4       0       1
12h          12          5       0       1
14h          14          6       0       1
24h          24          7       0       1
h2Control     2          8       1       0
h4Control     4          9       1       0
h8Control     8         10       1       0
h10Control   10         11       1       0
h12Control   12         12       1       0
h14Control   14         13       1       0
h24Control   24         14       1       0

> head(ncounts)
         2h 4h 8h 10h 12h 14h 24h h2Control h4Control h8Control h10Control h12Control h14Control h24Control
A1BG-AS1 14 17  8   4   4   7  15        17        17        17         17         17         17         17
A1CF      0  3  1   5   0   0   1         0         0         0          0          0          0          0
A2M      11 15 90  34  50  49  38         3         3         3          3          3          3          3
A2M-AS1   7  2 22   4   6  12   8         0         0         0          0          0          0          0
A2MP1    38 30 17  35  13  13   0         9         9         9          9          9          9          9
AAAS      6 59 83   4  67  97 150         1         1         1          1          1          1          1
> dim(ncounts)
[1] 165  14

Thanks a lot,

john

Hi,

I had the same error with the T.fit function.
In my case the error was due the column names of the design matrix starting with numbers instead of letters (e.g. 10wt instead of wt10).
I simply renamed the columns and this time it worked.

Hope it helps,

Paolo

Anyone else figure this out? I get in 1700 genes, then I get the same error. 

I also got the same error a few times, but never really figured out what caused it. Changing the step.method option from "backward" to "forward" did seem to solve the issue in one case.