Hi Alessandra and all.

I got this same problem: "Run-time error '1004': application-defined
or object-defined error" in Excel 2000 SP3.

Does upgrade to 2003 solved that problem?

Another question: there is a qBase version for Linux (Debian :) )? Or
a package for GNU R?

Thank you very much

Marcelo

--- In qBase@yahoogroups.com, Hayley McGrice <hayley.mcgrice@...> wrote:
>
> Hi Alessandra
>
> I experienced the same program when trying to run qBase with excel
> 2000.  Upgrading to excel 2003 fixed this problem straight away.
>
> Hope this helps
> Hayley
>

> > Hi
> >
> > I just downloaded qBase earlier this week, but I
> > am having trouble getting the program started. The
> > Excel program freezes everytime I try to analyze
> > an experiment, so I haven't been able to complete
> > the tutorial. I am also unablea to import my runs,
> > because I get the following message: "Run-time error
> > '1004': application-defined or object-defined error"
> >
> >Has anyone encountered this problem before? What can
> > I do to solve it?
> >
> > Thanks,
> >
> > Alessandra Splendore

Hi,

I have 2 questions:

1. I have samples with different quality I want to compare. I used 2
ref genes that are normally very stable in this tissue (we tested this
before), after calculation of the NF, it ranged from 0.1 to 6 between
the samples, illustrating clearly that there is a difference in
quality and/or quantity. Will it matter to use more reference genes
and is there even a reliable method to compare samples with a
different quality and/or quantity? These are the only samples I have
and it is difficult to collect new ones.

2. I have a gene I want to investigate that has a very low expression
in the tissue I have RNA from. In all samples the Cp value is around
34-38 (in non-diluted cDNA)? Is it still reliable to test expression-
differences? What is the maximum Cp value for a reliable analysis.

Kind regards

Hi, I have one more question about the inter run calibration performed
by qbase. If I decied to only include IRC's for one or two control
genes in my analysis wil qbase use these genes to perform inter run
calibration on all samples (i.e. all genes in an experiment that are
present on different runs with just the control IRC's) or will inter
run calibration only be performed on those genes that have IRC's
present?

Best Regards,

Damon

Hi I was wondering if it would be possible to do biological replicates
for each gene instead of doing 2 or 3 replicates of each sample. i.e.
give each biological replicate the same name and let qbase treat them
as per usual. I only have enough sample to assay each biological sample
in singlicate but have many samples per experimental group.

Is the statistical analysis applied by qbase compatible with this kind
of experimental design?

Thankyou,

Damon

Hi Jo, Hi Jans,

I'm trying to analyze some data in qBase but I'm experimenting a
problem (runtime error 1004) when I add more than 1 reference gene for
normalization... It seems to be in relation with the IRCs I use but I
don't be sure of this. I send you a qBase export file of the
experiment and the screen capture of the error in qBase.
Thanks for your help.

Regards
GG

Hi I was wondering if it would be possible to do biological replicates
for each gene instead of doing 2 or 3 replicates of each sample. i.e.
give each biological replicate the same name and let qbase treat them
as per usual. I only have enough sample to assay each biological sample
in singlicate but have many samples per experimental group.

Is the statistical analysis applied by qbase compatible with this kind
of experimental design?

Thankyou,

Damon

Hi, I'm really stuck with the IRC selections in qBase, hope you can
help me.

I have a large number of samples and genes spread across 7 runs.
Aside from the actual IRC, there are a number of samples tested on
different runs, with different genes. However, I repeated the HKGs
for those repeated samples on the different runs. eg.

Run 1: Sample 1: HKG1/2/3, Target gene1/2/3
Run 2: Sample 1: HKG1/2/3, Target gene4/5/6

qBase has thus picked up sample 1 as an IRC for HKG1/2/3 and renamed
the wells with HGKs accordingly, while leaving the sample name as
before for wells with target gene, leaving me with 3 samples ie.

"Sample 1_R1" (and _R2) has HKG data only
"Sample 1" has target gene data only

Thus normalisation factor calculations can not be selected.

Is it possible at all to undo IRC settings for a selected number of
samples only? or vice versa appoint a certain number of samples as
IRC specifically?


Kind regards

Hi there, I'm using LinReg for calculation of amplification efficiency
by averaging across all samples for each gene. Is there a way to
calculate the error propagation of the efficiency this way? Maybe from
the Pearson correlation values obtained during efficiency calculation?

Kind regards

Hi there, I'm using LinReg for calculation of amplification efficiency
by averaging across all samples for each gene. Is there a way to
calculate the error propagation of the efficiency this way? Maybe from
the Pearson correlation values obtained during efficiency calculation?


Kind regards

We are using samples from human blood which are stimulated with anti
CD3CD28.
When I tested several of these samples (using the similar
startconcentrations), I noticed that whereas the non-stimulated
samples have a normalisation factor around 1 (1.14; 1.07, 1.18 for 3
different samples) the NF for the stimulated samples is much higher
(2.24; 1.82; 2.13; 2.56) (using in the analysis 4 references genes).
Are these values still OK?
Is there a way I can reduce these values ?(unfortunately I cannot
adjust the stimulation protocol)

I tested with these samples several housekeeping genes. I found
strong differences in Ct values between the stimulated and non-
stimulated samples, so I did some standardcurves to evaluated the
reference gene quality (see table below)

CV M (geNorm)
B2M 19,56% 0,3866
EF1 9,94% 0,2942
UBC 11,25% 0,3122
YWHAZ 15,62% 0,3573
Mean 14,09% 0,3376

Is my conclusion correct that although I see these strong variations
in Ct values, that the reference genes themselves are actually OK?
Is EF-1 indeed the best housekeeping gene in this setup?

I hope you can help me with this,
Tania

We are using samples from human blood which are stimulated with anti
CD3CD28.
When I tested several of these samples (using the similar
startconcentrations), I noticed that whereas the non-stimulated
samples have a normalisation factor around 1 (1.14; 1.07, 1.18 for 3
different samples) the NF for the stimulated samples is much higher
(2.24; 1.82; 2.13; 2.56) (using in the analysis 4 references genes).
Are these values still OK?
Is there a way I can reduce these values ?(unfortunately I cannot
adjust the stimulation protocol)

I tested with these samples several housekeeping genes. I found
strong differences in Ct values between the stimulated and non-
stimulated samples, so I did some standardcurves to evaluated the
reference gene quality (see table below)

   CV M (geNorm)
B2M 19,56% 0,3866
EF1 9,94% 0,2942
UBC 11,25% 0,3122
YWHAZ 15,62% 0,3573
Mean 14,09% 0,3376


Is my conclusion correct that although I see these strong variations
in Ct values, that the reference genes themselves are actually OK?
Is EF-1 indeed the best housekeeping gene in this setup?

I hope you can help me with this,
Tania

Hi all,

what was the problem? I'm experiencing the same conflict of results...

Does having excluded samples in qBase matter?

Thanks
Ferdinando

--- In qBase@yahoogroups.com, jvdesomp <jvdesomp@...> wrote:
>
> This should not be the case. Please send your qBase export file and
> geNorm input file so we can have a look (qbase @ medgen dot ugent
dot be).
>
> Jo
>
> aperesbota schreef:
> >
> > Hi
> >
> > I have ranked my genes with geNORM but when I analyse with qBASE the
> > same set of data the M values and the evaluation of the reference
genes
> > is totally different. The CVs are very high (over 100 in some cases)
> > and the best M value calculated with qBASE is over 1....
> > I am doing probably something wrong....please advice.\
> >
> > Greetings,
> >
> > Adrian
> >
> >
>

I have done a real-time-pcr experiment with 2 house keeping genes
and two targetgenes of interest. I analysed the data with qBASE. The
reference gene quality evaluation gives the following values:


B2M
CV 24,16%
M 0,6866

UBC
CV 23,49%
M 0,6866


The paper (PMID: 17291332) states that mean CV and M values lower
than 25% and 0.5, respectively, are typically observed for stably
expressed reference genes in relatively homogeneous sample panels.
The CV values are below 25% but the M values are just above 0.5. Is
this good enough to use as reference genes?

Also, I compared the expression of the two reference genes between
the three groups I have by exporting the table from qbase into SPSS
(Mann withney and Kruskall Wallis test) and both were significant
between some of the three groups (not between all groups). These are
the GEOMAEN values per group for the two reference genes

Mean B2M
Mean UBC

B2M
Group1 1,400172192
Group2 1,578444476
Group3 1,945113663

UBC
Group1 1,921937419
Group2 1,70487044
Group3 1,383488985


This is a 0.71 less expression in group 1 compared to group 3 for
B2M, and a 1.38 higher expression of group 1 to group 3 for UBC. The
difference for both genes is in a diffent direction. The difference
is significant for group1 and group 3, and for group 2 and group 3
for both genes. Is it OK to use these reference genes? The target
genes are also significantly different between all groups, but with
a much higher difference in expression (factor 3-5 higher
expression) and a very significant p-value.

Kind regards

I have done a real-time-pcr experiment with 2 house keeping genes
and two targetgenes of interest. I analysed the data with qBASE. The
reference gene quality evaluation gives the following values:


B2M
  CV 24,16%
  M  0,6866

UBC
  CV 23,49%
  M 0,6866


The paper (PMID: 17291332) states that mean CV and M values lower
than 25% and 0.5, respectively, are typically observed for stably
expressed reference genes in relatively homogeneous sample panels.
The CV values are below 25% but the M values are just above 0.5. Is
this good enough to use as reference genes?


Also, I compared the expression of the two reference genes between
the three groups I have by exporting the table from qbase into SPSS
(Mann withney and Kruskall Wallis test) and both were significant
between some of the three groups (not between all groups). These are
the GEOMAEN values per group for the two reference genes




  Mean B2M
  Mean UBC

B2M
  Group1 1,400172192
  Group2 1,578444476
  Group3 1,945113663

UBC
  Group1 1,921937419
  Group2 1,70487044
  Group3 1,383488985



This is a 0.71 less expression in group 1 compared to group 3 for
B2M, and a 1.38 higher expression of group 1 to group 3 for UBC. The
difference for both genes is in a diffent direction. The difference
is significant for group1 and group 3, and for group 2 and group 3
for both genes. Is it OK to use these reference genes? The target
genes are also significantly different between all groups, but with
a much higher difference in expression (factor 3-5 higher
expression) and a very significant p-value.



Kind regards

Hi,

In my study i have run 2 housekeeping genes and 2 genes of interest.
Each gene was run on a 384 plate of cDNA samples.  My problem is that
several samples in each run failed to produce good replicates.  Is it
possible to repeat the failed samples on a new plate assuming that i
use inter-run calibrators?  The samples which need repeated are not
the same for each gene.

Regards
Robyn

I am using Cepheid and converting run files to qBase file format as
described in manual. However, I have a housekeeping gene tested in
each sample. Where does that Ct go? Do I make a separate run file or
list out each Ct for the housekeeper below the data? Thanks

I am using Cepheid and converting run files to qBase file format as
described in manual.  However, I have a housekeeping gene tested in
each sample.  Where does that Ct go?  Do I make a separate run file or
list out each Ct for the housekeeper below the data?  Thanks

I am running into a problem with the run editor view where it is
either empty (shows the grid, but no gene/sample names) or shows only
the upper left quadrant of my plate (with gene/sample names).

I use the run editor view to set which of my wells contain the
standards in order to create the standard curve. Is there any way
around this, and is anyone else having this problem?

I am using version 1.3.5.

Luke

Hi

When using the sample maximization method (all samples fit on one
plate/run but have many genes so have a number of plates)if you have a
repeat of the experiment using everything the same except a different
RT of you samples is it best to treat this as a new experiment in
qBase or can it be analysed within the same experiment and if so would
you require IRC?

Anton

Hi

When using the sample maximization method (all samples fit on one
plate/run but have many genes so have a number of plates)if you have a
repeat of the experiment using everything the same except a different
RT of you samples is it best to treat this as a new experiment in
qBase or can it be analysed within the same experiment and if so would
you require IRC?

Anton

> annafra07 <annafra07@...> ha scritto:  A: qBase@yahoogroups.com
> Da: "annafra07" <annafra07@...>
> Data: Mon, 06 Aug 2007 11:38:01 -0000
> Oggetto: [qBase] suddently analysis does not work
>
>         Dear Jo,
> I was using qBase successfully when it started to give errors in the
> raw data analysis (run-time 1004) (initialisation ok) , only with some
> excel import files apparently correct. I tried everything to get round
> the problem, without success. I am very frustrated, could you help me?
> Thanks in advance, Anna
>

This kind of problem is sometimes caused by gene or sample names that
are not recognized as text.  This can occur with names that are
interpreted as numbers or as dates.

Jan

>
>
>
>
>
> ---------------------------------
>
> ---------------------------------
> L'email della prossima generazione? Puoi averla con la nuova Yahoo! Mail
>

A: qBase@yahoogroups.com
Da: "annafra07" <annafra07@...>
Data: Mon, 06 Aug 2007 11:38:01 -0000
Oggetto: [qBase] suddently analysis does not work

Dear Jo,
I was using qBase successfully when it started to give errors in the
raw data analysis (run-time 1004) (initialisation ok) , only with some
excel import files apparently correct. I tried everything to get round
the problem, without success. I am very frustrated, could you help me?
Thanks in advance, Anna

--- In qBase@yahoogroups.com, "o_o_phil" <o_o_phil@...> wrote:
>
> --- In qBase@yahoogroups.com, "hellemans_jan" <jan.hellemans@> wrote:
> >
> > --- In qBase@yahoogroups.com, "o_o_phil" <o_o_phil@> wrote:
> > >
> > > Hello,
> > >
> > > Is the per-sample normalization factor (NF) calculated and applied
> > > independently for each run?
> > >
> > > For example, in the qBase paper (Genome Biology 2007 8(2):R19) there
> > > is an experimental setup example with 3 genes of interest and 3
> > > reference genes on 11 samples all in duplicates that spans on two
> > > plates (see paper Figure 2, run 4 and 5). The 3 ref genes are on the
> > > first plate, but not on the other. How does qBase calculate the NF
> > > factors for the 2nd plate? Does it uses the same NF calculated
on the
> > > 1st plate?
> >
> > Yes. The normalization factor (NF) is sample dependent.
> >
>
> Thank you for the quick reply!
>
> Ok I just want to see if I understand clearly. In the paper example
> illustrated in figure 2 there are two different plate layouts. The one
> with two plates (run 4 and 5) only need ref genes on one plate since
> all samples fit on it, whereas on the first layout ref genes are on
> all three plates to span all samples. Is that it? Would this mean that
> you need to place ref genes only once to cover all samples?

Exactly.


> In my case study I have 16 samples for about 40 GOIs and 2 refs
> (96-plates) The 16 samples are on columns and the GOIS on rows. We
> replicated the 2 refs on all plates. How does qBase handles replicates
> for refs?

It cannot deal with this properly. I suggest you use reference data
from only one plate. The alternative is to give the reference genes
slightly different names on the different plates (e.g. beta-actin-1,
beta-actin-2, ... for plate 1, 2, ...), after which qBase can use all
reference data (up to 5 reference genes).

Does it calculate and apply the NF for a sample
> independently for each plate, i.e. for the GOIs of that specific plate
> only, or does it average the NF for a sample for all replicates across
> all plates an apply it on all GOIs?
>
> I just want to verify if we get an advantage of replicating the refs
> on all plates or if we would get the same relative quantification
> results if we had only a single set of experiments for the refs to
> cover all 16 samples.

There is no real advantage in replicating reference genes. You only
need to quantify them once.  Remember, it is all about relative
quantification, whereby you compare samples (and not genes). If you
want to compare the expression levels between genes, you need
'absolute' standard curves (of which you known in advance the
concentration in terms of copies or number of molecules).

> thanks again for the insight!
>

--- In qBase@yahoogroups.com, "hellemans_jan" <jan.hellemans@...> wrote:
>
> --- In qBase@yahoogroups.com, "o_o_phil" <o_o_phil@> wrote:
> >
> > Hello,
> >
> > Is the per-sample normalization factor (NF) calculated and applied
> > independently for each run?
> >
> > For example, in the qBase paper (Genome Biology 2007 8(2):R19) there
> > is an experimental setup example with 3 genes of interest and 3
> > reference genes on 11 samples all in duplicates that spans on two
> > plates (see paper Figure 2, run 4 and 5). The 3 ref genes are on the
> > first plate, but not on the other. How does qBase calculate the NF
> > factors for the 2nd plate? Does it uses the same NF calculated on the
> > 1st plate?
>
> Yes. The normalization factor (NF) is sample dependent.
>

Thank you for the quick reply!

Ok I just want to see if I understand clearly. In the paper example
illustrated in figure 2 there are two different plate layouts. The one
with two plates (run 4 and 5) only need ref genes on one plate since
all samples fit on it, whereas on the first layout ref genes are on
all three plates to span all samples. Is that it? Would this mean that
you need to place ref genes only once to cover all samples?

In my case study I have 16 samples for about 40 GOIs and 2 refs
(96-plates) The 16 samples are on columns and the GOIS on rows. We
replicated the 2 refs on all plates. How does qBase handles replicates
for refs? Does it calculate and apply the NF for a sample
independently for each plate, i.e. for the GOIs of that specific plate
only, or does it average the NF for a sample for all replicates across
all plates an apply it on all GOIs?

I just want to verify if we get an advantage of replicating the refs
on all plates or if we would get the same relative quantification
results if we had only a single set of experiments for the refs to
cover all 16 samples.

thanks again for the insight!

--- In qBase@yahoogroups.com, "o_o_phil" <o_o_phil@...> wrote:
>
> Hello,
>
> Is the per-sample normalization factor (NF) calculated and applied
> independently for each run?
>
> For example, in the qBase paper (Genome Biology 2007 8(2):R19) there
> is an experimental setup example with 3 genes of interest and 3
> reference genes on 11 samples all in duplicates that spans on two
> plates (see paper Figure 2, run 4 and 5). The 3 ref genes are on the
> first plate, but not on the other. How does qBase calculate the NF
> factors for the 2nd plate? Does it uses the same NF calculated on the
> 1st plate?

Yes. The normalization factor (NF) is sample dependent.

>
> I have a second question about IRCs. Suppose we have a gene that spans
> 3 plates and one IRC, we have 3 CNRQ values for this genes. How does
> qBase select a final CNRQ value in that case, by averaging?

qBase does not calculate one finale CNRQ value, but three (one for
each plate).  If wanted, they can be averaged during the next steps of
your data processing.

>
>
> thank you for your help!
>
> Philippe
>

--- In qBase@yahoogroups.com, "o_o_phil" <o_o_phil@...> wrote:
>
> Hello,
>
> A quick one: what RQ calculation method should I use if I have a
> single reference gene?
>
> I see that if I use the qBase method I get normalization factor NF
> always at 1 ... should I revert to another method like the
> delta-delta-Ct method in this case?

When you have only 1 reference genes, the qBase quantification model
is automatically simplified to the delta-delta-Ct model with
efficiency correction (Pfaffl et al., 2001). If you are indicating
that the PCR efficiency of gene of interest and reference gene is
100%, the model is even further simplified to the delta-delta-Ct
method from Livak and Schmittgen (2001). In other words, the qBase
quantification model is universally applicable, and will equal Pfafll
or Livak/Schmittgen depending on conditions (one vs. multiple
reference genes; gene specific efficiency vs. 100% efficiency for all
genes).

>
> thanks!
>
> Philippe
>

Hello,

Is the per-sample normalization factor (NF) calculated and applied
independently for each run?

For example, in the qBase paper (Genome Biology 2007 8(2):R19) there
is an experimental setup example with 3 genes of interest and 3
reference genes on 11 samples all in duplicates that spans on two
plates (see paper Figure 2, run 4 and 5). The 3 ref genes are on the
first plate, but not on the other. How does qBase calculate the NF
factors for the 2nd plate? Does it uses the same NF calculated on the
1st plate?

I have a second question about IRCs. Suppose we have a gene that spans
3 plates and one IRC, we have 3 CNRQ values for this genes. How does
qBase select a final CNRQ value in that case, by averaging?


thank you for your help!

Philippe

Hello,

A quick one: what RQ calculation method should I use if I have a
single reference gene?

I see that if I use the qBase method I get normalization factor NF
always at 1 ... should I revert to another method like the
delta-delta-Ct method in this case?

thanks!

Philippe

Visual basic seems to halt in the line:

Workbooks(AnalyzerName).Worksheets("AnalyzerTemp").Range("B50") =
"Experiment"




--- In qBase@yahoogroups.com, "millasovovich" <sylvesterholt@...> wrote:
>
>
> I tried to load my expriment files from the stratagene Mx3000P, which
> qBASE should support. The loading of the files work fine, but I get
> "Runtime error 9 - subscript of range" when the samples are
initialised!?
>
> After this the program asks if I want to Debug. And if press no it
> gives this message:
> "An application-defined or object-defined error"
>
> Please help!
>
> best regards Sylvester
>