Automate dChip functions
Cheng Li
functions
--------------------------------------------------------------------------------
Dear Cheng,
Is it possible with dChips automation tools to automate a CNV analysis
and export the results without user interactions?
rshill
lch Posted: 31 Mar 2008 01:53 pm Post subject:
--------------------------------------------------------------------------------
You may try the latest dchip and the "Tools/Automate functions". You
can perform a menu step and add it there, and then select all the
added steps to run.
Cheng
lch Posted: 03 Aug 2007 12:43 pm Post subject: Automating dChip
functions
--------------------------------------------------------------------------------
RE: dChip and Windows Compute Cluster
Hi John,
Thanks for the nice illustration. You may use try this 8/11/07
version:
Http://biosun1.harvard.edu/~cli/dchip2006.exe
And see details here:
http://www.dchip.org/install.htm#automated
for some preliminary functions to automate dChip. Let me know any
issues arising and I can improve it.
Cheng
From: John McGee
Sent: Wednesday, August 08, 2007 4:12 PM
To: Cheng Li
Subject: RE: dChip and Windows Compute Cluster
Hi Cheng,
Thanks for the quick reply!
we are interested in running dChip on a Windows compute cluster by
partitioning the compute tasks. Diana needs to perform intensity based
copy number analysis on 700 subjects with 700 control cases. We plan
to partition the subjects into ten groups of 70, where each of the ten
groups will include all 700 control cases. We would then distribute
the ten separate tasks across the compute nodes. We see this as a
first step towards scaling the system further with significantly more
subjects, in addition to providing rapid iterations for the analysis.
Is this feature (intensity based copy number analysis) of dChip
available via command line so that we can kickoff the compute tasks
with scripts?
Best regards,
John McGee
John,
Thanks for the message. dChip doesn't support Compute Cluster since
it's a single machine program.
One possibility is you may divide CEL files into subsets (each
including a common normalization baseline array) and do parallel "Open
group" and "Normalization" function. However "Analysis/Model" needs to
be done using all arrays if "PM/MM difference model" is used. However
I think if array number is less than 500, running on one computer
overnight might be simpler.
Cheng
From: John McGee
Sent: Friday, August 03, 2007 10:27 AM
Subject: dChip and Windows Compute Cluster
Hello Cheng,
I work at the Renaissance Computing Institute at UNC Chapel Hill, and
one of our colleagues is interested in speeding/scaling up the
computations she makes using dChip.
I was wondering if you or anyone you know has looked into using dChip
with Windows Compute Cluster Services? We have a 20 core windows
compute cluster, which will likely grow significantly over the next
year, and are interested to see if there is a good fit with dChip.
Regards,
John McGee
Mon Oct 9, 2006 6:59 pm
--- In dc...@yahoogroups.com, "Cheng Li" <cli@...> wrote:
Hi Libi,
1) There is no automated way. However if you check "Open group/Perform
normalization afterwards", multiple steps will be run at once.
2) Affy right now recommends using their file SDK to read binary CEL
files.
Such codes are available at
http://www.affymetrix.com/support/developer/filesdk/index.affx?terms=no
Cheng
From: Libi Hertzberg [mailto:felibi@...]
Sent: Monday, October 09, 2006 6:22 AM
To: cli@...
Subject: dChip questions
Dear cheng,
I would like to implement a program that analyzes SNP chips data, and
I
would like to use dChip normalization method.
1) Is there a way to run the normalization for SNP chips in an
automated way
(without the need to go through the user interface)?
2) Can you direct me to information on how to read CEL files
(information on
what their structure is)? I was looking for such information but with
no
success in the meantime.
Thank you,
Libi
Cheng Li
Thanks for the question. Please try this 1/8 version:
Http://biosun1.harvard.edu/~cli/dchip.exe
In the INI file, specify
NORM_BASE_ARRAY=
as the array name (or the sample name if a sample info file is used) of the
baseline array.
Let me know how it goes.
Cheng
From: li...@mail.nih.gov [mailto:li...@mail.nih.gov]
Sent: Tuesday, January 06, 2009 2:29 PM
To: Cheng Li
Subject: Re: Automate dChip functions
Dear Dr. Li,
As I communicate with you earlier, I was trying to normalize
expression array with a specific array as baseline. Using your new
automate module, I am now able to create automate steps and run it in
command line as below:
C:\dChip>dchip_2008_05 -ini -norm -model test_auto.ini
I need to define a baseline array next. I have tried to edit the .ini
file by adding "BASELINE_ARRAY=HF1589_U133P2", but this is not
working. I also tried to use -baseline option in command line it does
not work also. Please direct me to achieve this step.
Thanks,
Aiguo Li
Cheng Li
To add each function in automation, you need to manually execute "Open
group" or "Normalize & Model" through menu; after each step, try to add it
to the automate step list.
Cheng
From: Li, Aiguo (NIH/NCI) [E] [mailto:li...@mail.nih.gov]
Sent: Friday, January 09, 2009 2:58 PM
To: Cheng Li
Subject: RE: Automate dChip functions
Dear Dr. Li,
With the new built when I build the automate module the NORM_BASE_ARRAY
is included in the .ini file and it does use the array I defined as
baseline array. However, when I go to Tool\Automate window, only the
ToolsExportExpression() module shows up. I could not add the
OpenGroup() and NormModel() modules into the automate window by clicking
the "+" sign as I did before with the previous build. Could you please
help me with this issue?
Thanks,
Aiguo