Agilent Feature Extractor Plug
Contents
1. Agilent Feature Extraction Plug In for GeneSpring GX Flag description Value in FE output GeneSpring flag Feature is saturated 1 Absent Feature is not uniform 0 Pass Feature is not positive and significant 0 Absent Feature is a population outlier 1 Marginal Feature is manually marked 0 Pass Background is not uniform 0 Pass Background reading is population outlier 0 Pass Is Control Probe 0 Pass Absent Table3 Feature is both saturated and a population outlier and gets flagged Absent since that is the lowest of the flag scores If the background information is used to determine the flag values those features that are marked as either having a non uniform background or are background population outliers get marked as M or Marginal probes by default or whatever is set in the dialog box table 4 Agilent Feature Extraction Plug In for GeneSpring GX 17 Flag description Value in FE output GeneSpring flag Feature is saturated 0 Pass Feature is not uniform 0 Pass Feature is not positive and significant 1 Pass Feature is a population outlier 0 Pass Feature is manually marked 0 Pass Background is not uniform 0 Pass Background reading is population outlier 1 Marginal Is Control Probe 0 Pass Marginal Table 4 Feature is a background population outlier and gets marked Marginal since this is the lowest of flag scores 18 Agilent Feature Extraction Plug In for GeneSpring GX Appendix B Agilent File F2. genes were functioning normally when sampled and that their intensities are normal for what they were doing By calling that normal state 1 for every chip in the experiment you can normalize these variations in gene intesities These normalizations are only made on genes considered normal and flagged Present Genes flagged Marginal or Absent are ignored to avoid skewing the data Per gene Normalize to the median The last step in normalizing Agilent FE 1 Color data involves normalizing the expression value for each gene across all arrays e For Agilent FE 1 Color data set the cutoff value to 0 01 in Raw Signal data measurement values and click OK Agilent Feature Extraction Plug In for GeneSpring GX 23 amp Per Gene Normalize to median Normalize each gene to the median of tle measurements for that gene Cutoff ifthe median is below the cutoff the cutoff value will be used instead See Help for details Cutomvalue 1 0E 6 in TM measurement values Apply only to Specific Samples OK Cancel Help There are two options that can be set in this normalization step 1 a cutoff value can be set to ensure that values lower than the cutoff will be replaced with the cutoff value and 2 the cutoff value can be applied to either the normalized value or the Raw signal value The recommended normalization step has a cutoff set for the Raw signal value of 1 0E This together with the cutoff value for the Control v
3. lens 30 Data nereerutes Ireveforr wer gt beer s Det Tn seeretee Ove aren Pa tpl Cem Cy cone i harsa Dota Trncteerutes Sersvo ccats moans Por Sent ord Pe CNE iera deot ort cet Fer CUP MOPEDS N Sopota pioto Pa Cap Newnalze x oesB rebel gres T ihe Pot r dt Tent Bees Pa Cup Newnan a cornteteder Her Ore Hore nee D cose Gennes Uae a Oped renare Seve At Convio Pa Ore Noreere h mesan anng Por Cap ane Per Gere Moans panning No warnnge 2 cee wo Agilent Feature Extraction Plug In for GeneSpring GX The default two color data normalization scenario is the Per Spot and Per Chip Intensity dependent Lowess normalization Since Agilent Feature Extraction already applies Lowess normalization to the data there is no need for a second Lowess normalization technically however it would not be detrimental For Agilent FE 2 color data a more appropriate normalization scenario is to simply apply a ratio calculation of the two channels followed by a per gene normalization to ensure all ratio s for each gene are normalized to 1 During installation of the plug in a Saved Scenario was installed to perform this normalization To apply the normalization scenario 1 Select Use a Saved Scenario Select a Normalization Scenario No Norrmalizations Standard One Color Standard Two Color Load Scenario Delete Scenario Rename Scenario 2 Select the Agilent FE or Agilent FE 1 color scenario from the lis
4. the Feature Extraction output file and stores them as Sample Attributes in GeneSpring GX The fields that are extracted from the FE output file are located in line 2 of the FE output file marked FEPARAMS and in line 6 marked STATS ED Microsoft Excel U5 14702384_1601152 021078_502_AD1 txt a October 1AE vernon A E EJE tot fost integer fox fost float rtoge aias SDak OSsetSrd0 qOark OtserDak Ofse DarcOthe Dart Ofre ark v7 tM 1000 1667 7 a45 ango rtoger ten lon tort woga vtoge Row Cal Saits Prot GesBack GerPege Probe Contr 9 AAPG AAFS 2A AES AALSSEB3 AAC22032 Cabins AACIITI Mee Whus14702304_ 16011521021078 E YY Resty All items from the FEPARAMS are extracted as Sample Attributes The names of the Sample Attributes are the same as the name of the FEPARAMS or STATS fields 30 Agilent Feature Extraction Plug In for GeneSpring GX Agilent Feature Extraction Plug In for GeneSpring GX 31 www agilent com In this book This book contains brief instructions to help you get started with your Agilent Feature Extractor Plug In for GeneSpring GX The information includes instruction on e Installing the Feature Extraction Plug In e Loading and normalizing data produced by Feature Extraction Software e Identifying flags related to specific types of problems Agilent Technologies Inc 2005 Printed in USA First Edition October 2005 l Agilent Technologies
5. Agilent Feature Extractor Plug In for GeneSpring GX FE Plug In Quick Start Guide This document describes the new Feature Extractor Plug In for GeneSpring GX What is the Feature Extractor Plug In The plug in enhances the importing of data files produced by Agilent Feature Extraction software The plug in allows you to import a different set of columns from a FE data file and a number of FEPARAM fields as Sample Attributes and to convert Agilent flag information into the GeneSpring GX flag information The source code for the plug in is available and is located in the source directory of the Programs folder in the GeneSpring data folder It is released under the LGPL license What is in this Quick Start guide e Installation of the Plug In e Appendix A Flags e Appendix B Agilent File Format e Appendix C Agilent Normalization Scenario for 1 Color Data e Appendix D Agilent Normalization Scenario for 2 Color Data e Appendix E Extracted Sample Attributes ee Agilent Technologies Installation of the Plug In The FE plug in for GeneSpring GX can be installed easily by dragging and dropping the Installer ZIP file onto an open main window of GeneSpring GX The file is a special GeneSpring ZIP file that allows the easy installation of the new functionality into GeneSpring GX 1 Download the installer ZIP file from the Agilent Technologies Web site at http www chem agilent com script
6. Feature Extraction Plug In for GeneSpring GX Appendix C Agilent Normalization Scenario for 1 Color Data For optimal use of the expertise that is built into Agilent Feature Extraction software Agilent recommends that you use a normalization scenario that is different than the default one color scenario This recommended normalization scenario makes use of the converted flag information that allows you to exclude the control and otherwise flagged probes The recommended Agilent FE 1 Color normalization involves three steps The Agilent FE 1 Color normalization scenario has three steps 1 Data transformation Set measurements from lt 0 0 to 0 0 2 Per chip Normalize to the median or percentile 3 Per gene Normalize to median Each step is discussed in detail Experiment Normalizations Demonstration Experiment Order of Normalizations to Perform Add Normalization Step gt Inspect a Data Transformation Set measurements less than 0 01 to 0 01 Delete em Per Chip Normalize to 50th percentile Start with pre normalized values 3 Per Gene Normalize to median M Data Transformation SAGE transform Move Down Data Transformation Real Time PCR transform Data Transformation Subtract background based on negative controls Use Defaults Data Transformation Set measurements less than 0 01 to 0 01 Data Transformation Transform from log to linear values Data Transformation Dye swap Per Spot Divide by control channel Data Tran
7. alue of 1 0E ensures that all normalized values are sensible Normalization completed When you have completed the three steps for normalizing Agilent FE 1 Color data the following will have occurred e Values below 0 01 were set to 0 01 e Each measurement was divided by the 50 0th percentile of all measurements in that sample e Each gene was divided by the median of its measurements in all samples If the median of the raw values was below 10 then each measurement for that gene was divided by 10 if the numerator was above 10 otherwise the measurement was thrown out Agilent Feature Extraction Plug In for GeneSpring GX Appendix D Agilent Normalization Scenario for 2 Color Data For optimal use of the expertise that is built into Agilent Feature Extraction software Agilent recommends that you use a normalization scenario that is different than the default two color scenario usually used This recommended normalization scenario also makes use of the converted flag information that allows you to exclude the control and otherwise flagged probes The recommended normalization involves three steps 1 Per spot Divide by the control channel 2 Per chip Normalize to the 50th percentile 3 Per gene Normalize to the median Each step will be discussed in detail S Experiment Normalizations Demonstration Experiment Order of Normalizations to Perform 14 Per Spot Dice by control channel Per Chip Normalize to 50th pe
8. ay contain more than one data format and it will always contain the Custom data format GeneSpring GX can recognize many different data file formats but it will Agilent Feature Extraction Plug In for GeneSpring GX 5 NOTE To ensure that the FE plug in will be used be sure that the Choose File Format drop down box is set to Agilent or Agilent FE 1 Color as appropriate to the data type 7 only report the names of data files that are consistent with the current data files If the Agilent FE output file contains 2 color data the format is called Agilent If the file contains 1 color data the format is called Agilent FE 1 Color More than the two data formats Agilent and Custom may be available in the drop down menu if you or the administrator have created a Recognized Custom format before This Recognized Custom Format will most likely be chosen as the default Verify that the Select Genome field indicates the correct genome for loading the data files and press Next to continue The Import Data Selected Files dialog box displays to allow you to add more files to be imported You can now choose multiple files to import by navigating to the correct folder on the left Select the files in the middle window and press Add to add the files to the list on the right Press Next to continue import sta Sebected Fies NOTE Ifyou want to attach image files to Samples such as JPG images of the arrays you must p
9. default Agilent Feature Extraction Plug In for GeneSpring GX 7 No Yes Import only FE flags that deal with background information Click Advanced if you want to import only specific FE flags The Advanced Flag Import dialog box displays with a more granular set of choices See step 12 Flag Import Do you wart GeneSpring to use the spot information in FE to flag the data ves Do you wart to include the background readings in the flag settings No be Lox _ advanced _Heto_ The three buttons at the bottom of the dialog box offer the options e OK continues the import process by displaying the Suggested Normalization window and starting to read and convert the data files e Advanced displays the Advanced Flag Import dialog box where you can change the default conversion settings e Help displays the appropriate Help page URL 12 Click Advanced to access the Advanced Flag Import dialog box where you can change the conversion settings Agilent Feature Extraction Plug In for GeneSpring GX Advanced Flag Import You can respond Yes or No to the prompt Do you want GeneSpring to use the spot information in FE to flag the data e Ifyou respond Yes the first five choices in the Spot Problems box are enabled along with the Mark all control probes as Absent check box e If you respond No the entire Spot Problems box is disabled along with the Do you want to include the background readings in the flag se
10. ed probes 28 To include the features that are only flagged to be marginal the option can be changed by selecting Present or Marginal from the drop down list Per gene Normalize to median The last normalization step normalizes the expression value for each gene across all the arrays around 1 This allows you to compare the genes across all the conditions or arrays regardless of the actual normalized expression value ratio Per Gene Normalize to median Normalize each gene to the median of the measurements for that gene Cutoff Ifthe median is below the cutoff the cutoff value will be used instead See Help for details CutotValue 1 0E 6 Raw Signa _ measurement values Apply only to Specific Samples ok Cancet Help There are two options that can be set in this normalization step 1 a cutoff value can be set to ensure that values lower than the cutoff will be replaced with the cutoff value and 2 the cutoff value can be applied to either the normalized value or the Raw signal value Agilent Feature Extraction Plug In for GeneSpring GX The recommended normalization step has a cutoff set for the Raw signal value of 1 0E This together with the cutoff value for the Control value of 1 0E ensures that all normalized values are sensible Agilent Feature Extraction Plug In for GeneSpring GX 29 Appendix E Extracted Sample Attributes The Agilent FE plug in extracts a number of fields from
11. er chip normalizations control the intensity of chip wide variations These variations may result from inconsistent washing inconsistent sample preparation or other microarray production or microfluidics imperfections The Normalize to a Median or Percentile option lets you divide all of the measurements on each chip by a specified percentile value By default this value is 50 0 e For Agilent FE 1 Color data type 50 0 in the percentile text box if the default value is not already displayed Agilent Feature Extraction Plug In for GeneSpring GX e Check the Use only measurements flagged box and set the drop down menu to Present Only e Click the Never apply extra background correction button Click OK when done S Per Chip Normalize to a median or percentile This step is appropriate for large arrays when most of he genes are unaffected by experimental parameters Normalize each chip to the 50 th Percentile of the measurements taken from that chip Restrict Measurements used in the Cakulation of the Percertiie F Use only measurements flagged EET I Use only measurements with Background Correction Never apply extra background correction Always apply extra background correction C needed apply extra background correction I Apply oniy to Spectic Samples OK Conve Heip Identical chips from different samples may show different intensities due to external factors You can offset this by asuming that all
12. ever there a number of features in GeneSpring GX that cannot use this flag information unless the information is first converted into special GeneSpring flags The Agilent FE plug in converts the complex set of FE flags into three levels of GeneSpring flags Absent A Marginal M or Present P The flag columns from Agilent FE output and their meaning are provided in table 1 Flag description Column headers in Agilent FE output Feature is saturated glsSaturated rlsSaturated Feature is not uniform glsFeatNonUnifOL rlsFeatNonUnifOL Feature is not positive and significant glsPosAndSignif rlsPosAdSignif Feature is a population outlier glsFeatPopnOL rlsFeatPopnOL Feature is manually marked IsManualFlag Background is not uniform glsBGNonUnifOL rlsBGNonUnifOL Background reading is population outlier glsBGPopnOL rlsBGPopnOL Is Control Probe Control Type Table 1 Flag columns in Agilent FE output The g and r used as the first letter of the column header name indicate the Green Cy3 and Red Cy5 channels respectively Flag logic To map the complex 14 binary flag columns there are 4096 flag combinations to the simple three value GeneSpring flags A M and P a mapping that depends on the seriousness of the flags was applied Certain flag values carry more weight that other flag values The most serious flags are mapped as Absent medium serious flags are mapped as Marginal and all other features are mapped as P
13. lace the image files and data files in the same folder amas 4 1601162000054 a AN A DNONGSE APA PORN _ 100149000 9294_ 8x a_i eADNNTET_ T Sorghes hee mtoe tes 540 Es After you select all of the files that you want to load the Import Data Preprocess Data Files dialog box displays Agilent Feature Extraction Plug In for GeneSpring GX Import Data Preprocess Data Files Do you want to preprocess your data files before importing your samples GTi ENR AE ETRE DREN Previous Next Cancel Help 10 Select Enhanced Agilent FE Import from the drop down menu and click Next Import Data Preprocess Data Files Do you want to preprocess your data files before importing your samples None 7 Enhanced Agilent FE Import 11 The plug in imports the correct columns for example for 2 color data rProcessedSignal is imported as SIGNAL and gProcessedSignal as CONTROL and it also allows you to convert Agilent flags to GeneSpring flags If you choose to use the Enhanced Agilent FE Import plug in as directed in step 10 the Agilent FE flag information is converted to GeneSpring flag information The Flag Import dialog box displays with two questions How you answer the questions determines how GeneSpring GX will interpret the FE flags e No No Ignore all of the FE flags e Yes Yes Import all of the FE flags e Yes No Import only FE flags that deal with spot information This is the
14. malization is not skewed by outliers or control probes It also ensures that normalized values are sensible amp Per Chip Normalize to a median or percentile This step is appropriate for large arrays when most of the genes are unaffected by experimental parameters Normalize each chip tothe 50 th Percentile of the measurements taken tom that chip Restict Measurements used in the Cakulation of the Percentile Use only measurements lagged EI Use only measurements with Background Correction Never apply extra background correction Always apply extra background correction F needed apply extra background correction I Apply oniy to Spectic Samples OK Cancel Help The recommended scenario only considers probes that are marked with a Present or P GeneSpring flag This is indicated by the Normalization step option Use only measurements flagged Present Only Preprocessing the data using the Agilent FE plug in s default settings results in the exclusion of probes that are control probes saturated feature population outliers non uniform features or manually flagged Agilent Feature Extraction Plug In for GeneSpring GX 27 NOTE The features are only excluded in calculating the median value The features are not excluded from the experiment and continue to be available for review or analysis The excluded features will also have the same normalization applied to when compared to the non exclud
15. nome from the Agilent Web site by selecting File Agilent Feature Extraction Plug In for GeneSpring GX 3 gt Import Genome See the User Manual for more information on importing or creating genomes in GeneSpring GX S Ful GeneSpring Human 1A Genes Fee Edt View Experimerts Colorter C Gene Lists C Experiments CI Gene Trees CI Condition Trees Ctassiticatons Patrways CI Aray Layouts CJ Expression Proties CJ Externa Programs CJ Bootrnarks TJ Secrets all genes Fitering Tools Aractstons Window Help Expression Magrefcabon 3 Select File gt Import data 4 In the Select One Data File window select an FE output file to import from the file chooser dialog box provided Agilent Feature Extraction Plug In for GeneSpring GX 5 The Import Data Define File Format and Genome window displays Import Data Define File Format and Genome File Format Choose File Format M Genome Select the genome set of genes on the array for this data If your genome does not appear on the list you can create a new one by selecting Create a New Genome Select Genome E Genomes or Arrays Agilent EN Demo Chips 3 Demo Human S Demo Rat gt Yeast C Create a New Genome Next Cancel Help 6 GeneSpring GX tries to determine the data file s format is by examining the contents If it is successful it reports the name in the Choose File Format drop down box The drop down menu m
16. oi _ciose Hee For 2 color data Suggested Normaiizations for Agilent aa 0 x 15 Here the import procedure is the same as the normal import procedure Continue to import the samples Propert Data Sample Attributes Agilent Feature Extraction Plug In for GeneSpring GX 11 Agilent FE normalization scenarios When all samples are loaded and used to build an experiment we recommend that you apply the Agilent FE or Agilent FE 1 Color scenario New Experiment Checklist You are almost finished creating your experiment Before you begin analysis you should set up its normalizations experimental parameters and error model and choose your default experiment interpretation You may reach these windows using the buttons below Altematively you may find them in the Experiments menu gt New Experiment Checklist Define Normalizations Normalizations Define Parameters Parameters Define the Default Interpretation Experiment Interpretation Define the Error Model Error Model To apply the normalization scenario click the Normalizations button in the New Experiment Checklist dialog box or click Experiments gt Experiment Normalization S Eapermment Mormatzatioss Artery Ves w Riteta a Apal avs Per Coto eter et cope vent d ramm corveiedinn errr tr Deta Transteereten GACT rretra Date Tameeerutes Soa Tore PCA versa Dea NAP T E ET baIeE Data Trensicerwton Set essaenerts
17. ormat Agilent FE transfers these fields to GeneSpring GX GeneSpring information Agilent FE fields For 2 color data For 1 color data Gene Name Signal Reference a k a Control within GeneSpring User Provided Signal Precision Description GenBank ID Flags Sample Attributes ProbeName If column is empty use SystematicName If that column is empty use FeatureNum rProcessedSignal gProcessedSignal LogRatioError is converted using the following calculation For a given gene in the FE output let RPS rProcessedSignal imported as Signal GPS gProcessedSignal imported as Control LRE logRatioError Then import the following signalprecision GPs e 1 Description GenBank The preprocessor needs to translate the eight different flag settings All of the FEPARAMS and STATS fields are loaded as Sample Attributes The names of the Sample Attributes are the names provided on the FEPARAMS or STATS row The values are the values displayed below the field name in the DATA row The type in the TYPE row can be used to determine if a Sample Attribute is numeric or not 1 FE type integer or float GeneSpring numeric 2 FE type text gt GeneSpring non numeric Same gProcessedSignal N A gProcessedSigError Same Same Same Same Table 5 Interpreting the Agilent file format Agilent Feature Extraction Plug In for GeneSpring GX 19 20 Agilent
18. rcentie Start with pre normalized values Per Gene Normalize to median Oata Transformation SAGE transform Data Transformation Real Time PCR transform Data Transformation Subtract background based Data Transformation Set measurements less Mar Oata Transformation Transform from log to linear x Data Transformation Oye swap Per Spot Divide by contro channel Data Transformation Reserve contol channel Per Spot and Per Chip intons y dependens Lowes I Use Recommended Order Got Text Description Per Crip Normalize to a median or percentile Per Chip Normalize to postive contol genes Use a Saved Scenario Saves Scenario Per Chip Normalize to a constand value Wamings Per Gene Normalize to specific samples Per Gene Normalize to median Per Chip and Per Gene Median polishing 24 cancer Hein Per spot Divide by control channel GeneSpring GX extracts two columns from the Agilent FE output file that represent the red and green channel signal values after processing in Agilent Feature Extraction These two channels Agilent Feature Extraction Plug In for GeneSpring GX 25 called rProcessedSignal for the red or Cy5 channel data and gProcessedSignal for the green or Cy3 channel data are loaded respectively as the Signal and Control channels in GeneSpring GX Any normalization scenario in GeneSpring GX involves the ratio between the Signal and Control channels The first step in the normalization scenario is to create
19. resent Agilent Feature Extraction Plug In for GeneSpring GX 15 Table 2 16 Flag values in the FE output are independent flags that is a feature can be saturated but uniform and a background population outlier GeneSpring GX will keep the Lowest GeneSpring flag value low is A next is M then followed by P For example if a feature is not flagged as either saturated or non uniform but flagged as a population outlier and not manually marked or a control probe and not making use of the Background mapping it will have a score of P P M P P P P and will therefore receive the GeneSpring flag M or Marginal table 2 Flag description Value in FE output GeneSpring flag Feature is saturated 0 Pass Feature is not uniform 0 Pass Feature is not positive and significant 1 Pass Feature is a population outlier 1 Marginal Feature is manually marked 0 Pass Background is not uniform 0 Pass Background reading is population outlier 0 Pass Is Control Probe 0 Pass Result Marginal Feature is a population outlier and gets marked Marginal since this is the lowest of flag scores A second example concerns a feature that is flagged as a saturated feature and as a population outlier and all other flags are not set in FE output It gets marked as A because the feature is saturated Since this is the lowest possible flag value in GeneSpring GX the entire feature is flagged as A or Absent table 3
20. s generic asp page 34734 amp indcol Y amp prodcol Y 2 Save the ZIP file to your local hard drive 3 3Start GeneSpring GX 4 4Drag the ZIP file onto the main GeneSpring GX window 5 A dialog box displays Install Enhanced Agilent FE Import This will install the Enhanced Agilent FE Import Plug in GeneSpring will quit when this operation 1s finshed Do you wish to proceed 6 Press OK to continue or Cancel to cancel the installation process 2 Agilent Feature Extraction Plug In for GeneSpring GX 7 A Caution dialog box may display indicating that the installation will overwrite certain files This may happen if previously the plug in had been installed or a normalization scenario with the name Agilent FE was created You can abort the installation by pressing No and determine if the files can be safely overwritten Caution This upgrade will overwrite some existing files which were most recently installed on Apr 14 2005 11 05 01 AM Do you wish to proceed 8 After the installation is complete a new dialog box displays indicating that the installation was successful Upgrade Complete Upgrade complete Click Ok to quit GeneSpring 9 Press OK to continue 10 GeneSpring GX closes automatically The plug in is now installed correctly and Agilent FE data can be loaded Loading data produced by Agilent FE 1 Start GeneSpring GX 2 Ifyou do not yet have the Agilent genome in GeneSpring GX download the ge
21. sformation Reserve control channel Per Spot and Per Chip Intensity dependent Lowess normalization Per Chip Normalize to a median or percentile Per Chip Normalize to positive control genes I Use Recommended Order Get Text Description Per Chip Normalize to a constant value z x Use a Saved Scenario Save As Scenario Per Gene Normalize to specific samples Aami Per Gene Normalize to median amn Per Chip and Per Gene Median polishing z ES OK Cancel Help Agilent Feature Extraction Plug In for GeneSpring GX 21 22 Data transformation Set measurements from lt 0 01 to 0 01 The data transforming normalization involves applying a mathematical modification to a variables values During normalization GeneSpring GX recalculates the data values and uses them in subsequent analyses The Set Measurements option sets any measurements less than a specified cutoff value to the cutoff value By default this value is 0 01 e For Agilent FE 1 Color data set the cutoff value to the default by typing Cutoff 0 01 in Cutoff text box and click OK Data Transformation Set measurements less than 0 01 to DER This sets values below the cutoff to the cutoff Cutoff 0 01 l Apply only to Specific Samples OK Cancel Help By default this cutoff value is set to 10 but for this example scenario we have lowered this value to 0 01 to allow lower values Per chip Normalize to the median or a percentile P
22. t and press Load Scenario See appendix B for details on the Agilent FE Normalization scenario Agilent Feature Extraction Plug In for GeneSpring GX 13 3 The new normalization scenario is loaded and displayed in the Normalization window Ceperment Normatizations Artery Vein w R tbets Data aratere ater Ret mreaserereres ees Rar 6 on te OUT Date Tueratorr ation Tsin fore wo D liver sires Date Tureferration Oye mne Pr Sor Conde gt corbu benret Cata Turari Reseree reat oriant Per Spy anc Pm Cap Firat Jepentert owent Normeti Por Chp A and Ted por t For Chat ROPv ante V pe Ame erru peros Pe Crog Mcrae 0 e coe ttentvaue Per Gene Nomisi te mark tampes Per Cerne Nanaia o melas Fer Crag ad Por Gone Matisa oo biring 4 Press OK to apply the new normalization scenario to the experiment Normalizing Experiment DOR Please wait while Artery Vein w IL1beta is normalized This may take some time Loading Data 2 samples loaded After the Agilent FE normalization scenario is applied the data is ready to be analyzed in GeneSpring GX in the usual manner Agilent Feature Extraction Plug In for GeneSpring GX Appendix A Flags Agilent FE output data contain a large number of flags that can be used to indicate various problems with a feature on an array Usually this flag information can be easily used in GeneSpring GX after the data files are imported using the Filter on Data file filter How
23. this ratio with the normalization step Per Spot Divide by control channel amp Per Spot Divide by control channel This divides the Signal Channel by the Control Channel This is recommended for two color experiments if you do not use the intensity dependent normalization Cutoff ifthe Control Channel is very low a cutoff value will be used instead See Help for details Cutoff 1066 I Apply only to Specific Samples The Control Channel cutoff value is the only option that you can set for this normalization step If the Control Channel value is below the cutoff value the cutoff value is used for the control channel to ensure that sensible ratios are calculated By default this cutoff value is set to 10 but for this example scenario we have lowered this value to 1 0E to allow lower values for the control channel Agilent Feature Extraction Plug In for GeneSpring GX Per chip Normalize to 50th percentile The second normalization step normalizes the expression value to the median of the expression values on the array This ensures that the expression values for each array can be compared across chips by centering each of the distributions for each array around 1 making comparison across arrays possible When you determine the median value for the distribution of signal values you can exclude from the calculation each probe that has been flagged as being unreliable This ensures that the median calculation and nor
24. tting prompt The new data file that the plug in creates will contain a GeneSpring Flags column where everything is marked P for Present You can also respond Yes or No to the prompt Do you want to include the background readings in the flag settings Agilent Feature Extraction Plug In for GeneSpring GX 9 If you respond No the last two choices in the Spot Problems box are disabled Background is not uniform and Background reading is a population outlier and both options are imported as Present If you respond Yes the last two choices are enabled and set to Marginal Alternatively you can change the conversion setting for each FE flag value by selecting the appropriate radio buttons and pressing OK to begin importing See appendix A for a detailed explanation of the flag logic and how the settings in the Advanced Flag Import dialog box influence the outcome 13 A window displaying a progress bar displays indicating that the preprocessing is active Preprocessing Data Files DOR Preprocessing Files Enhanced Agilent FE Import 14 The Suggested Normalizations for Agilent FE dialog box displays automatically after you have completed the Flag Import settings The version of this dialog box that displays depends on whether the plug in recognized the file as being in 1 color or 2 color file format 10 Agilent Feature Extraction Plug In for GeneSpring GX For 1 color data E Suggested Normalizations for Agilent ig i
Download Pdf Manuals
Related Search