[BioC] flowCore: inverse logicle transformation of flow cytometry data
Nishant Gopalakrishnan
ngopalak at fhcrc.org
Thu Oct 8 08:38:39 CEST 2009
Hi Pyne,
The inverse function is available in flowCore (1.11.22 ) and makes use of a
general biexponential function for its calculation. You can find more
information about the function and an some example code on how to use it by
typing in
library(flowCore)
? inverseLogicleTransform
Regarding your question regarding the spread of points around zero, once you
have fixed the maximum value for your display scale (r) and the width of the
display (d), the parameter w controls the strength of linearization around
zero.
So in short you will want to adjust parameter w.
Nishant
-----Original Message-----
From: spyne at broadinstitute.org [mailto:spyne at broadinstitute.org]
Sent: Wednesday, October 07, 2009 4:52 PM
To: ngopalak at fhcrc.org
Subject: RE: [BioC] flowCore: inverse logicle transformation of flow
cytometry data
Hi Nishant,
Could you please tell me the usage (or the call) for the method
that you wrote - inverseLogicleTransform - I guess there is
no documentation yet - hence I thought I might ask you.
Also, sorry to ask again, could you please tell me exactly which
parameter in the logicle function determines the rate of spread
of the points away from zero (origin)?
Thanks a lot for your co-operation.
-Pyne
Quoting Nishant Gopalakrishnan <ngopalak at fhcrc.org>:
> Hi Pyne,
>
> I have checked in a method inverseLogicleTransform which you can use to
get
> back to the original untransformed data provided you pass in the same
> parameters w,r,d which were used to transform the data initially. This
> should be available using biocLite("flowCore") tomorrow afternoon.
>
> The default values for the parameters of the logicle transform were
obtained
> from the paper by Parks et al. referenced in the man page for the
function.
> For detailed information regarding the effect of the change in these
> parameters on the logicle function and how to select optimal parameters
for
> your data under consideration, please refer to the paper by Parks et al.
> which goes through the details of selection of the width parameter,
dynamic
> range etc.
>
> Nishant
>
> -----Original Message-----
> From: spyne at broadinstitute.org [mailto:spyne at broadinstitute.org]
> Sent: Tuesday, October 06, 2009 1:15 PM
> To: Nishant Gopalakrishnan
> Subject: Re: [BioC] flowCore: inverse logicle transformation of flow
> cytometry data
>
>
> Hi Nishant,
>
> Thanks a lot for your reply.
>
> The inverse function will be quite helpful for me
> and to many other users who may want to go back to
> the scale they are used to once the analysis is over.
>
> By the way, the logicle function has about 5 parameters,
> I guess. However, I cannot tell this clearly from the
> function description, between w, r, and d, exactly which
> parameter determines the rate of spread of points away
> from zero (origin)? And how was its default value set?
>
> Best regards,
> -Pyne
>
>
> Quoting Nishant Gopalakrishnan <ngopalak at fhcrc.org>:
>
>> Hi Pyne,
>>
>> I am working on a function to calculate the inverse and will be checking
>> in some changes today.
>> Thanks in advance for your patience.
>>
>> Nishant
>>
>> spyne at broadinstitute.org wrote:
>>>
>>> Hi,
>>>
>>> Wondering if the inverse function of logicle was implemented.
>>>
>>> Thanks,
>>> -Pyne
>>>
>>>
>>> Quoting Chao-Jen Wong <cwon2 at fhcrc.org>:
>>>
>>>> Hi, Pyne,
>>>>
>>>> I agree with you that it is good to have an inverse function. Thanks
for
>>>> your suggestion and tips. We will try to implement it next week.
>>>>
>>>> Thanks,
>>>> Chao-Jen
>>>>
>>>> spyne at broadinstitute.org wrote:
>>>>>
>>>>> Hi,
>>>>>
>>>>> The reason I need the inverse function for logicle is because
>>>>> after I have computationally identified the cluster of events in
>>>>> logicle-transformed marker space, now I want to use the knowledge
>>>>> of that range of events in the original, untransformed scale for
>>>>> sorting out similar events in the subsequent experiments.
>>>>>
>>>>> My guess is that this may not be a very far-fetched scenario,
>>>>> and since the transformation is deterministic and bijective
>>>>> anyway, an inverse function would be good to have, at least for
>>>>> the default argument settings. One option is of course a slow
>>>>> numerical computation method.
>>>>>
>>>>> However, since the transformation is monotonic, for a fixed setting
>>>>> of arguments (e.g. the default setting), doing a simple binary search
>>>>> over a reasonable range is a cheap way to approximate the inverse
>>>>> within a desirable accuracy.
>>>>>
>>>>> Thanks!
>>>>> -Pyne
>>>>>
>>>>>
>>>>> Quoting Chao-Jen Wong <cwon2 at fhcrc.org>:
>>>>>
>>>>>> Hi, Pyne
>>>>>>
>>>>>> That is an interesting question. flowCore does not have an inverse
>>>>>> function for the logicle transformation. Since the logicle
>>>>>> transformation is an one-to-one and onto function, it is possible to
>>>>>> implement an inverse function. It is, however, not
>>>>>> straightforward. Do
>>>>>> you really really need such a function?
>>>>>>
>>>>>> spyne at broadinstitute.org wrote:
>>>>>>>
>>>>>>> Hi,
>>>>>>>
>>>>>>> I applied logicle transformation (with default arguments)
>>>>>>> to my data points, then detected the subpopulations of
>>>>>>> interest in the transformed data, and now I want to
>>>>>>> revert the subpopulations back to the original scale of
>>>>>>> the untransformed state.
>>>>>>>
>>>>>>> In other words, if I want to apply the inverse of the logicle
>>>>>>> transformtion (applied with default arguments, which I do not
>>>>>>> know) to my data, is that possible?
>>>>>>>
>>>>>>> Thanks.
>>>>>>> -Pyne
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> 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
>>>>>>
>>>>>>
>>>>>> --
>>>>>> Chao-Jen Wong
>>>>>> Program in Computational Biology
>>>>>> Division of Public Health Sciences
>>>>>> Fred Hutchinson Cancer Research Center
>>>>>> 1100 Fairview Avenue N., M2-B876
>>>>>> PO Box 19024
>>>>>> Seattle, WA 98109
>>>>>> 206.667.4485
>>>>>> cwon2 at fhcrc.org
>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>>> --
>>>> Chao-Jen Wong
>>>> Program in Computational Biology
>>>> Division of Public Health Sciences
>>>> Fred Hutchinson Cancer Research Center
>>>> 1100 Fairview Avenue N., M2-B876
>>>> PO Box 19024
>>>> Seattle, WA 98109
>>>> 206.667.4485
>>>> cwon2 at fhcrc.org
>>>>
>>>>
>>>
>>> _______________________________________________
>>> 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
>>
>>
>
>
>
>
More information about the Bioconductor
mailing list