FlowCap - Revision history

Keilwagen at 08:59, 21 February 2013

2013-02-21T08:59:46Z

← Older revision		Revision as of 08:59, 21 February 2013
Line 1:		Line 1:
	This page contains supplemental information to our submission to Dream6 challenge 4 "Classification of AML". For information about this challenge, please visit [http://www.the-dream-project.org/challenges/dream6flowcap2-molecular-classification-acute-myeloid-leukaemia-challenge the Dream6 homepage]. In the following we describe our approach and provide a download of the program used to make the predictions for the challenge and its source code.		This page contains supplemental information to our submission to Dream6 challenge 4 "Classification of AML". For information about this challenge, please visit [http://www.the-dream-project.org/challenges/dream6flowcap2-molecular-classification-acute-myeloid-leukaemia-challenge the Dream6 homepage]. In the following we describe our approach and provide a download of the program used to make the predictions for the challenge and its source code.

			== Paper ==
			The paper '''''Critical assessment of automated flow cytometry data analysis techniques''''' has been published in [http://www.nature.com/nmeth/journal/vaop/ncurrent/full/nmeth.2365.html Nature Methods].

	== Method description ==		== Method description ==

Grau: /* Instructions */

2012-04-27T14:26:08Z

Instructions

← Older revision		Revision as of 14:26, 27 April 2012
Line 76:		Line 76:
	you obtain the predictions we submitted to the DREAM challenge.		you obtain the predictions we submitted to the DREAM challenge.

	If you want to use this method for other data sets, feel free to adapt the sources to ~~our~~ specific problem. To compile the sources, you also need [[Downloads \| Jstacs 2.0]] (sources or binary .jar) in your class path.		If you want to use this method for other data sets, feel free to adapt the sources to your specific problem. To compile the sources, you also need [[Downloads \| Jstacs 2.0]] (sources or binary .jar) in your class path.

Grau: /* Instructions */

2012-04-27T14:25:52Z

Instructions

← Older revision		Revision as of 14:25, 27 April 2012
Line 68:		Line 68:
	load ... load (load the classifier (instead of storing it), default = false) = false		load ... load (load the classifier (instead of storing it), default = false) = false
	threads ... compute threads (the number of threads that are use to evaluate the objective function and its gradient, valid range = [1, 128], default = 2) = 2		threads ... compute threads (the number of threads that are use to evaluate the objective function and its gradient, valid range = [1, 128], default = 2) = 2


			If Dream6C4.jar is started from a directory containing
			* the description file 'DREAM6_AML_TrainingSet.csv' of the challenge,
			* a directory 'csv' with the CSVs of all patients and filenames of the format 'XXX.CSV', where XXX is equal to the file number (with additional leading 0s) in 'DREAM6_AML_FilesTubesAndSubjects.CSV', and
			* the classifier description 'final-classifiers.xml', which is part of the binary download and can also be downloaded seperately
			you obtain the predictions we submitted to the DREAM challenge.

			If you want to use this method for other data sets, feel free to adapt the sources to our specific problem. To compile the sources, you also need [[Downloads \| Jstacs 2.0]] (sources or binary .jar) in your class path.

Grau at 09:44, 25 April 2012

2012-04-25T09:44:27Z

← Older revision		Revision as of 09:44, 25 April 2012
Line 56:		Line 56:
	* [http://www.jstacs.de/downloads/Dream6-sources.zip Sources], additionally require Jstacs 2.0 sources to compile		* [http://www.jstacs.de/downloads/Dream6-sources.zip Sources], additionally require Jstacs 2.0 sources to compile
	* [http://www.jstacs.de/downloads/final-classifier.xml XML of the classifier] used in the challenge		* [http://www.jstacs.de/downloads/final-classifier.xml XML of the classifier] used in the challenge

			== Instructions ==
			You can start the JAR contained in the binary on the command line by calling
			java -jar Dream6C4.jar

			Optional parameters are
			home ... home directory (the path to the data directory, default = ./) = ./
			desc ... description file (file containing the description of patients for the training data, default = DREAM6_AML_TrainingSet.csv) = DREAM6_AML_TrainingSet.csv
			key ... unhealthy key (key in the description file that indicates unhealthy patient, all other keys treated as healthy, default = AML) = AML
			classifier ... classifier (path to the classifier; if load=true, this classifier is loaded from the file, otherwise it is stored to that file, default = final-classifiers.xml) = final-classifiers.xml
			load ... load (load the classifier (instead of storing it), default = false) = false
			threads ... compute threads (the number of threads that are use to evaluate the objective function and its gradient, valid range = [1, 128], default = 2) = 2

Grau: /* Download */

2012-04-25T09:38:48Z

Download

← Older revision		Revision as of 09:38, 25 April 2012
Line 53:		Line 53:

	== Download ==		== Download ==
	~~available soon~~		* [http://www.jstacs.de/downloads/Dream6-binary.zip Binaries], including the XML of the classifier used in the challenge
			* [http://www.jstacs.de/downloads/Dream6-sources.zip Sources], additionally require Jstacs 2.0 sources to compile
			* [http://www.jstacs.de/downloads/final-classifier.xml XML of the classifier] used in the challenge

Grau: /* Method description */

2012-04-24T09:44:32Z

Method description

← Older revision		Revision as of 09:44, 24 April 2012
Line 9:		Line 9:

	Following these assumptions we take a step-wise approach using Jstacs:		Following these assumptions we take a step-wise approach using Jstacs:

	1. We build a classifier that returns the probability that a specific cell from a specific tube is infected or not. Such a classifier is learned for each kind of tube (that means each selection of markers) independently, where the measurements of all cells of AML patients are used as foreground (positive) and the measurements of all cells of healthy patients are used as background (negative). The log-values of the measurements are modeled by normal distributions and the parameters are learned by the maximum conditional likelihood principle.		1. We build a classifier that returns the probability that a specific cell from a specific tube is infected or not. Such a classifier is learned for each kind of tube (that means each selection of markers) independently, where the measurements of all cells of AML patients are used as foreground (positive) and the measurements of all cells of healthy patients are used as background (negative). The log-values of the measurements are modeled by normal distributions and the parameters are learned by the maximum conditional likelihood principle.

	2. For each patient, we compute the fraction of cells in a tube classified as infected. For each patient we obtain a series of 8 such fractions, one for each tube.		2. For each patient, we compute the fraction of cells in a tube classified as infected. For each patient we obtain a series of 8 such fractions, one for each tube.

	3. We create another classifier working on these 8 values using a logistic regression and learn its parameters by the maximum supervised posterior principle based on the labeling of patients. The output of this classifier is the final prediction.		3. We create another classifier working on these 8 values using a logistic regression and learn its parameters by the maximum supervised posterior principle based on the labeling of patients. The output of this classifier is the final prediction.

Grau: Created page with "This page contains supplemental information to our submission to Dream6 challenge 4 "Classification of AML". For information about this challenge, please visit [http://www.the-dr..."

2012-04-24T09:10:54Z

Created page with "This page contains supplemental information to our submission to Dream6 challenge 4 "Classification of AML". For information about this challenge, please visit [http://www.the-dr..."

New page

This page contains supplemental information to our submission to Dream6 challenge 4 "Classification of AML". For information about this challenge, please visit [http://www.the-dream-project.org/challenges/dream6flowcap2-molecular-classification-acute-myeloid-leukaemia-challenge the Dream6 homepage]. In the following we describe our approach and provide a download of the program used to make the predictions for the challenge and its source code.

== Method description ==

We base our classifier on the following assumptions:
* each experiment (tube) is an independent indication if this patient suffers from AML or not
* for each cell in a tube we may independently decide if this cell is infected or not
* the number of cells classified as infected differs substantially between patients suffering from AML and healthy patients

Following these assumptions we take a step-wise approach using Jstacs:
1. We build a classifier that returns the probability that a specific cell from a specific tube is infected or not. Such a classifier is learned for each kind of tube (that means each selection of markers) independently, where the measurements of all cells of AML patients are used as foreground (positive) and the measurements of all cells of healthy patients are used as background (negative). The log-values of the measurements are modeled by normal distributions and the parameters are learned by the maximum conditional likelihood principle.
2. For each patient, we compute the fraction of cells in a tube classified as infected. For each patient we obtain a series of 8 such fractions, one for each tube.
3. We create another classifier working on these 8 values using a logistic regression and learn its parameters by the maximum supervised posterior principle based on the labeling of patients. The output of this classifier is the final prediction.

For the predictions for the unlabeled data, we use the trained classifiers and follow the protocol as before: Classify each cell in each tube - compute the fraction of cells classified as infected - use logistic function for final prediction.

This approach can be summarized by the following pseudo code:

== Pseudo code ==
=== Parameter learning ===
For each tube do
Load the measurements for each cell in this tube;
Compute log-values for all measurements;
Create a sample based on the log-values of the individual cells and label all cell stemming from AML patients as foreground class and all cells from healthy patients as background class;
Create a classifier based on 7 independent normal distributions, corresponding to the measurements for the two scatter and five antibody measurements, for the foreground and background class each;
Estimate the parameters of the classifier (i.e., means and standard deviations) from this sample using the maximum conditional likelihood (MCL) learning principle.
Classify the log-measurements of all cells of a patient and compute the fraction of cells (later denoted as patient posterior) with a probability P(AML | cell) > 0.5;
Done;

For each patient do
Create a sequence of the 8 patient posteriors;
Done;
Create a sample from these sequences with labels according to the patients’ state of health;
Create a classifier based on logistic regression;
Estimate the parameters of the classifier from this sample using the maximum supervised posterior (MSP) learning principle and a product normal prior with standard deviation 1;

=== Prediction ===
For each tube do
Load the measurements for each cell in this tube;
Compute log-values for all measurements;
Create a sample based on the log-values of the individual cells;
Classify the log-measurements of all cells of a patient using the previously trained classifier for this tube and compute the fraction of cells (later denoted as patient posterior) with a probability P(AML | cell) > 0.5;
Done;

For each patient do
Create a sequence of the 8 patient posteriors;
Done;

Finally, use the classifier based on logistic regression to obtain the final prediction based on this sequence of patient posteriors.

== Download ==
available soon