[BioC] Affy PA calls

[Loren Engrav]

Ok, sounds as if p and q trump PA calls

I will cease and desist on thinking to use PA calls after p and q to further
refine things

Thank you all


> From: Mark Cowley <m.cowley at garvan.org.au>
> Date: Fri, 30 Jul 2010 13:43:59 +1000
> To: Loren Engrav <engrav at u.washington.edu>
> Subject: Re: [BioC] Affy PA calls
> 
> no worries.
> i usually prioritise genes to follow up by small q-values and large logFC, but
> if there's a low Pcount or a low average expression level, then that puts me
> off too
> 
> mark
> On 30/07/2010, at 1:36 PM, Loren Engrav wrote:
> 
>> Sorry, on the last email I hit your address rather than the BioC address
>> 
>> 
>>> From: Loren Engrav <engrav at u.washington.edu>
>>> Date: Thu, 29 Jul 2010 19:53:00 -0700
>>> To: Mark Cowley <m.cowley at garvan.org.au>
>>> Conversation: [BioC] Affy PA calls
>>> Subject: Re: [BioC] Affy PA calls
>>> 
>>> Thank you
>>> Ok, I will not cut further on the PA strings as described below
>>> 
>>> But then as I look at the 171 probes of statistical interest without this PA
>>> cut, I see a probe that is P on only one of 60 chips.
>>> 
>>> In keeping with the "decision" that PA calls are too blunt to use to remove
>>> probes, I suppose I should ignore that as well, again thinking that p and q
>>> trump PA calls.  True?
>>> 
>>> 
>>> 
>>> 
>>> 
>>> 
>>>> From: Mark Cowley <m.cowley at garvan.org.au>
>>>> Date: Fri, 30 Jul 2010 12:43:06 +1000
>>>> To: Loren Engrav <engrav at u.washington.edu>
>>>> Cc: rbioc <bioconductor at stat.math.ethz.ch>
>>>> Subject: Re: [BioC] Affy PA calls
>>>> 
>>>> Hi Loren,
>>>> the mas5calls procedure generates Pvalues, and a call of Present just means
>>>> P
>>>> < threshold.
>>>> a PAP might mean 3 pvalues that are VERY similar to each other, and the 2nd
>>>> is 
>>>> slightly higher than the threshold
>>>> 
>>>> a PAP also may mean that your gene is close to the limit of detection on
>>>> the
>>>> array, and that's always a dangerous place to be on a microarray.
>>>> I think the p/values from MAANOVA will be more robust than
>>>> over-interpreting
>>>> P/M/A calls, since it incorporates the within group measurement variance
>>>> 
>>>> cheers,
>>>> mark
>>>> 
>>>> On 30/07/2010, at 12:29 PM, Loren Engrav wrote:
>>>> 
>>>>> Thank you both
>>>>> 
>>>>> Will try to explain better
>>>>> 
>>>>> Have Affy chips from shallow and deep wounds in 3 Duroc and 3 Yorkshire
>>>>> pigs
>>>>> 
>>>>> Prior to maanova and q-value I did strip all probes that were Absent on
>>>>> all
>>>>> 60 chips
>>>>> 
>>>>> Now then with maanova and q-value I have 171 probes of statistical
>>>>> interest,
>>>>> but
>>>>> 
>>>>> In this set of 171 statistically selected probes, the detection string for
>>>>> a
>>>>> given probe might be PAPAAP meaning
>>>>> 
>>>>> Shallow pig1 Present
>>>>> Deep pig1 Absent
>>>>> 
>>>>> Shallow pig2 Present
>>>>> Deep pig2 Absent
>>>>> 
>>>>> Shallow pig3 Absent
>>>>> Deep pig3 Present
>>>>> 
>>>>> With these PA calls, pig3 is different.  I think I would like to remove
>>>>> this
>>>>> probe from further consideration because even though it achieved p and q
>>>>> significance, the pigs are not the same. Or are the PA calls so "blunt"
>>>>> that
>>>>> they do not really indicate anything about shallow and deep, and after
>>>>> stripping those absent on all 60 chips, are best ignored?
>>>>> 
>>>>> Thank you again both
>>>>> 
>>>>> 
>>>>> From: Sean Davis <sdavis2 at mail.nih.gov>
>>>>> Date: Thu, 29 Jul 2010 20:37:25 -0400
>>>>> To: Loren Engrav <engrav at u.washington.edu>
>>>>> Cc: rbioc <bioconductor at stat.math.ethz.ch>
>>>>> Subject: Re: [BioC] Affy PA calls
>>>>> 
>>>>> 
>>>>> 
>>>>> On Thu, Jul 29, 2010 at 8:28 PM, Loren Engrav <engrav at u.washington.edu>
>>>>> wrote:
>>>>>> 
>>>>>> Sorry if was too terse
>>>>>> 
>>>>>> I have Affy chips from shallow and deep wounds in 3 Duroc and 3 Yorkshire
>>>>>> pigs. And I have the p-values from maanova and q-values from Storey
>>>>>> q-values
>>>>>> which returned 171 probes of interest.
>>>>>> 
>>>>>> I would also like the 3 Durocs and the 3 Yorkshire bio reps to be the
>>>>>> "same".
>>>>>> 
>>>>>> So the general question is how to define "same"?
>>>>>> 
>>>>>> The specific question is can the Affy PA calls be used to define same?
>>>>>> If
>>>>>> one pig is PA meaning shallow is Present and deep is Absent, can I say
>>>>>> the
>>>>>> other two pigs should also be PA?  Or is the PA call too suspect to be
>>>>>> used
>>>>>> in this way?
>>>>>> 
>>>>> 
>>>>> That is a pretty blunt instrument.
>>>>> 
>>>>> Doesn't your hypothesis test define sameness, at least relative to the
>>>>> other
>>>>> group?  Again, you may need to elaborate a bit more on what you are doing.
>>>>> 
>>>>> Sean
>>>>> 
>>>>> 
>>>>>>>> From: Mark Cowley <m.cowley at garvan.org.au>
>>>>>>>> Date: Fri, 30 Jul 2010 10:02:56 +1000
>>>>>>>> To: Loren Engrav <engrav at u.washington.edu>
>>>>>>>> Cc: rbioc <bioconductor at stat.math.ethz.ch>
>>>>>>>> Subject: Re: [BioC] Affy PA calls
>>>>>>>> 
>>>>>>>> hi Loren,
>>>>>>>> please elaborate
>>>>>>>> mark
>>>>>>>> 
>>>>>>>> On 30/07/2010, at 8:55 AM, Loren Engrav wrote:
>>>>>>>> 
>>>>>>>>>> Does anyone put much stock in Affy PA calls?
>>>>>>>>>> 
>>>>>>>>>> I searched gmane and I think the answer is No. True?
>>>>>>>>>> 
>>>>>>>>>> Thank you.
>>>>>>>>>> 
>>>>>>>>>> _______________________________________________
>>>>>>>>>> Bioconductor mailing list
>>>>>>>>>> Bioconductor at stat.math.ethz.ch
>>>>>>>>>> https://stat.ethz.ch/mailman/listinfo/bioconductor
>>>>>>>>>> Search the archives:
>>>>>>>>>> http://news.gmane.org/gmane.science.biology.informatics.conductor
>>>>>>>> 
>>>>>> 
>>>>>> _______________________________________________
>>>>>> Bioconductor mailing list
>>>>>> Bioconductor at stat.math.ethz.ch
>>>>>> https://stat.ethz.ch/mailman/listinfo/bioconductor
>>>>>> Search the archives:
>>>>>> http://news.gmane.org/gmane.science.biology.informatics.conductor
>>>>> 
>>>>> 
>>>>> 
>>>>> 
>>>>> [[alternative HTML version deleted]]
>>>>> 
>>>>> _______________________________________________
>>>>> Bioconductor mailing list
>>>>> Bioconductor at stat.math.ethz.ch
>>>>> https://stat.ethz.ch/mailman/listinfo/bioconductor
>>>>> Search the archives:
>>>>> http://news.gmane.org/gmane.science.biology.informatics.conductor
>>>> 
>> 
>> 
>