"user manual"
Contents
1. Final CNVR1 regions CNVR2 CNVR3 In st ep 1 since the pair consisting of CNV2 and CNV3 has the highest RO these two CNVs are merged into a CNV element called CNV2 3 Similarly in step 2 CNV2 3 and CNV4 are merged into CNV2 3 4 As the RO values of all the remaining pairs do not pass the RO threshold three CNV regions are defined in black The RO method can reduce the extent of size overestimation of CNV regions caused by CNVR method However compared with the other two methods the RO method may increase the possibility of false negative results For example suppose that one locus embedded within CNVR3 in the final step figure above is truly associated with the trait of interest In both CNVR and fragment methods the frequency of this locus is 2 but in the RO method the frequency is 1 which may cause the true association to be statistically missed 13 3 Fragment The fragment method dissects overlapping regions which have different frequencies of CNVs from the neighboring regions into smaller separate fragments CNVs _ e EEE ee CNVR1 CNVR2 CNVR3 CNVR4 CNVR5 Fragments Therefore this method has the least probability of size overestimation compared with the other two methods Also the potential of false negative results may be lower than the RO method By dissecting overlapping CNVs into smaller fragments this method could generate a large number of smaller CNV elements which may increa2. 33 on ejo RO_8 5 653 461 5 850 821 Gain RO_9 7 127 027 7 522 678 Loss Y Y Y Y Y 7 Y RO_10 x 1 z Y Y Y o 2 lt o N RI lt lt RO_11 61 723 356 530 RO_12 1 235 658 564 621 Loss RO_13 RO_14 RO_15 RO_16 RO_17 RO_18 RO_19 RO_20 RO_21 q za aaeeeo uan 1 229 608 235 657 ain RO 23 148 953 9 7 eE 1 235 658 356 530 Mixed RO 24 p___ 48 847 6 149 051 9 Loss 356 531 564 621 PTF loss RO 25 1149 202 8 Loss E 149 202 866 16 830 808 16 935 995 Gain 63 704 937 63 810 371 Gain 82 461 630 82 644 795 104 307 2 Gain 1 61 723 tess ean 1 85 924 Gain CNVR_14 l CNVR_17 1 CNVR_18 l CNVR_19 1 CNVR_20 1 ele 144 849 5 Gain 1 228 695 229 063 145 398 1 1 229 064 229 607 rey Fe ey Cy ed ed ed E E Piel Gain Gain N nN Di aa n 1 16 968 362 A 1 17 029 580 145 899 Gain 1 17 035 208 89 133 112 89 500 461 Gain 1 17 036 531 1 17 037 085 A 1 17 045 446 olojaojolajolajaojajaja 2 ziz 2 SIS lt ral pa 9 2 0 Ho Pro Pro Pro Pro Fra Pro ro ro ro RIo olojo lujo ma win Nie RO 26 mE Gai i 182 611 6 Loss 1 16 830 808 16 935 995 N N N RO_28 RO_29 RO_30 RO_31 RO_32 N NINN jee B Association analysis Report This is the window for the output
3. 1 740 857 1 030 307 Gain 1 16 830 808 16 935 995 Gain 1 16 968 362 17 298 496 Gain 1 63 704 937 63 810 371 Gain 1 82 461 630 82 644 795 Gain 1 104 130 168 104 307 231 Gain 1 121 343 784 121 482 967 Gain 1 144 036 737 144 849 544 Gain 1 145 206 610 145 398 179 Gain 1 148 530 424 149 051 903 Mixed 1 149 086 173 149 202 866 Loss 1 166 574 788 166 966 828 Gain 1 182 454 823 182 611 606 Loss 1 196 706 260 196 812 518 Gain 1 243 163 830 243 274 530 Gain 15 2 RO O bh bh bob b bo o bo hs bo db bd db db bd bh bb bo 3 Fragment O gt bh h h h h ob o bobo o bobo bd bh db db dh dh bh dh dh hb dh dh dh dh dh hh bh dh hh bh dh dh db db o Start 61 723 235 658 740 857 16 830 808 16 968 362 63 704 937 82 461 630 104 130 168 121 343 784 144 036 737 145 206 610 148 530 424 148 530 424 148 947 698 149 086 173 166 574 788 182 454 823 196 706 260 243 163 830 Start 61 723 85 924 228 695 229 064 229 608 235 658 356 531 740 857 16 830 808 16 968 362 17 029 580 17 035 208 17 036 531 17 037 085 17 045 446 17 177 034 17 182 426 17 190 851 17 245 519 63 704 937 82 461 630 104 130 168 121 343 784 144 036 737 145 206 610 148 530 424 148 662 752 148 947 698 148 953 985 149 086 173 149 086 551 149 190 307 166 574 788 182 454 823 196 706 260 196 711 067 243 163 830 End 356 530 564 621 1 030 307 16 935 995 17 298 496 63 810 371 82 644 795 104 307 231 121 482 967 144 849 544 145 398 179 14
4. Running After selecting statistical methods and setting allele frequency threshold level click Run key Then the output of statistical calculation will be displayed in the report screen 18 Report Screen A CNVR Report The CNV region determined by user s preference will be displayed in this window The same list is also stored as a tab delimited text file on the same directory The file name consists of original name and region type These are example tables for three different type of region CNVR RO Fragment El example ci GC cmt0x1s0c0x3 table O example ci GC rmt0x1s50c0x3 table Chromoso e CNVR ID Chromoso Start End Type j RO_O 2 985 939 3 086 364 Gain a RO 1 Y 3 352 707 3 538 704 Gain FRAG 2 Y 3 352 707 13 423 827 RO_2 3 613 770 13 784 346 FRAG_3 Y 3 423 828 3 471 481 RO_3 4 181 582 4 324 808 Gain RO 4 4 410 751 4 550 549 Gain ROS 4 670 883 14 780 229 Gain RO 6 4 909 027 5 013 153 Gain G Y RO 7 Yo 5 363 369 5 471 500 Y ziz OLIOL OLJOLOLO SSSSSS i ha OOS A Us lt lt lt 4 909 027 5 013 153 Gain 5 363 369 5 471 500 Gain 5 653 461 5 850 821 Gain 7 127 027 7 522 678 Loss 7 524 092 7 697 801 Loss 61 723 564 621 mixed 16 830 808 16 935 995 Gain 82 461 630 82 644 795 Gain i Gain 9 Gain Gain 145 206 6 1145 398 1 Gain 148 530 4 1149 051 9 mixed 149 086 1 1149 202 8 Loss
5. Places Properties Extraction Wizard Select a Destination Files inside the ZIF archive will be extracted to the location you choose Selecta folder to extract files to Files wall be extracted to this directory Browse Password Extracting CO Step3 Double click to execute CNVRuler exe CidownloadiCHYRuler File Edit View Favorites Tools Help Back a ya Search Wey Folders Heb Address C download CNyRuler Type File and Folder Tasks BIN File ar ChVWRuler exe Application gt CNWRulerManual pdf Adobe Acrobat Doc E readme txt Text Document Make a new Folder d Publish this Folder to the Web EJ Share this folder Other Places man download My Documents fm Shared Documents NOTICE Do not put programs nor data on the folder which has a name with 2 byte character ex Asian characters It makes R occurs inside error cannot make temporary directory After finishing installation user interface will be appeared Clinical Information CNY Region Association Test CNVRuler ver 1 3beta Clinical infomation file Sample ID Gender Covariate ONY call file Remove smallerthan LOD Bird Suite segment mean cutoff TOGA NimbleScan Method enve gt Recurrence _ Gain Loss separated regions Include CNVs across low frequency areas Method Use PCA as covariates for Population Stratification No w Minor Allele
6. association CNV Ruler can read 10 types of CNV call outputs see Table below and a custom CNV call Format PennCNV Nexus Genomic Workbench CGHscape TCGA files NimbleScan Genome Studio QuantiSNP BirdSuite Genotying Console Version Tested 2011Jun16 Ref Wang et al 2007 www biodiscovery com www agilent com Jeong et al 2008 cancergenome nih gov www nimblegen com www illumina com Colella et al 2007 2008 Korn et al www affymetrix com If you want to use your own CNV list file you must prepare a simple tab delimited text file containing 5 columns as the example below The names and order of column headings should be Chr Start End Event and Sample_ID respectively Example of user own CNV data file CAT Start Ena Event Sample ID al 10430 10592 Loss Syndrome TypeA O01 al 12410 12900 Loss Syndrome TypeA 01 2 400 6210 Garn Syndrome Typen 01 dl 2430 2592 Loss Syndrome TypeA 02 Filtering options There are two filtering options in the CNV data uploading section CNV call file Remove CNVs smaller size than segment mean cutoff TEGA NimbleScan 1 2 CNV size filter Users can set their own threshold for minimum size to define the CNVs unit bp A CNV which is smaller than the threshold will be excluded Mean signal intensity of the segment filter This option will be only used when the input file is TCGA or NimbleScan data TCGA and NimbleScan data do not have gai
7. 8 662 751 148 953 984 149 051 903 149 202 866 166 966 828 182 611 606 196 812 518 243 274 530 End 85 923 228 694 229 063 229 607 235 657 356 530 564 621 1 030 307 16 935 995 17 029 579 17 035 207 17 036 530 17 037 084 17 045 445 17 177 033 17 182 425 17 190 850 17 245 518 17 298 496 63 810 371 82 644 795 104 307 231 121 482 967 144 849 544 145 398 179 148 662 751 148 947 697 148 953 984 149 051 903 149 086 550 149 190 306 149 202 866 166 966 828 182 611 606 196 711 066 196 812 518 243 274 530 16 Type Mixed Loss Gain Gain Gain Gain Gain Gain Gain Gain Gain Gain Gain Loss Loss Gain Loss Gain Gain Type Mixed Mixed Mixed Gain Gain Mixed Loss Gain Gain Gain Gain Gain Gain Gain Gain Gain Gain Gain Gain Gain Gain Gain Gain Gain Gain Gain Gain Mixed Loss Loss Loss Loss Gain Loss Gain Gain Gain C CNVR Phenotype association analysis Association Test Method Lise PCA as covariates for Population Stratification No w Minor Allele Threshold eo Methods Logistic regression B Linear regression C Chi Squared D Fisher s Exact Test gt Users select one of the methods above Regarding the Chi Squared test users can select between Chi Squared test or Chi squared test with Yates continuity correction based on the characteristics of their data Logistic Regression Logistic Regression Linear Regression Chi Square Test Ch
8. CNVRuler User Manual v1 2 CNVRuler software is freely available with associated files and user manual in our website http www ircgp com CNVRuler index html Contact to developer Yeun Jun Chung yejun catholic ac kr and Ji Hong Kim lomolith gmail com CNV Ruler is designed for CNVR based association analysis with user friendly graphic interface All forms of major CNV call outputs from different segmentation tools such as Genotyping Console Genome Studio Genomic Workbench BirdSuite PennCNV and Nexus can be processed without additional converting steps CNV Ruler supports defining three different types of CNV regions CNVRs and four statistical methods for CNVR based association analysis Users can analyze CNVR phenotype associations with their preferable segmentation tools and can test various CNVR definitions and statistical methods suitable for their own study design 1 Prerequisites CNV Ruler needs Java Run time Environment of SUN Microsystems or equivalent JRE 1 6 0 or higher For all statistical analyses R is used as a calculation core o JRE If your system does not have Java Virtual Machine JVM you can download it from Oracle s Java home page http www oracle com technetwork j ava avase downloads index html For checking whether JVM is properly installed type java version on a terminal prompt The version of your JVM will appear if it is correctly installed Windows users can type the command on th
9. LINK a tool set for whole genome association and population based linkage analyses Am J Hum Genet 81 559 575 Subirana et al 2011 CNVassoc Association analysis of CNV data using R BMC Medical Genomics 4 47 The Cancer Genome Atlas TGCA research Network 2008 Comprehensive genomic characterization defines human glioblastoma genes and core pathways Nature 455 1061 1068 Wang K et al 2007 PennCNV An integrated hidden Markov model designed for high resolution copy number variation detection in whole genome SNP genotyping data Genome Res 17 1665 Wittig M et al 2010 CNVineta a data mining tool for large case control copy number variation datasets Bioinformatics 26 2208 2209 20
10. Threshold C Separated p values forGain Loss LRT 3 Data analysis A Data uploading Clinical Information Clinical infomation file Sample ID Phenotype Covariate CNV call file Remove smallerthan For CNV Ruler analysis two types of information Clinical and CNV data must be prepared Step 1 Uploading clinical data In the clinical information Cl file 4 items sample ID age sex and phenotype are to be included as separate columns in the Cl txt file see the example below lf age or sex information is not available the users can do the association analysis with just sample ID and phenotype data Phenotype means the dependent variable for regression analysis After selecting the Cl file for the analysis you must choose the sample ID and phenotype columns in the user interface with other covariates The sample ID should be matched to the name of the samples in the CNV data file Phenotype status must have binary values 0 and 1 for logistic regression For sex column users can input values as a string male or female m or f man or woman 1 or 0 1 or 2 and it is not case sensitive In addition to the four basic Cls other variables for logistic regression analysis can be added in your Cl see an example below If you have more Cls than the 4 Cl columns click the Covariates button then Covariates pop up window will appear You can select the extra
11. e command line window from Start button e R CNV Ruler needs R for its calculation process You can download it from its project home page http www r project org After selecting download mirror site and OS platform you can download the distribution binaries If your system doesn t have the R package on it CNV Ruler will prompt it and try to open the R download site NOTICE If CNV Ruler keeps warning that there is no R package after installation you may add PATH variable manually Test by typing R version on your terminal Usually Linux users do not need to change it 2 Installation The CNV Ruler package consists of two executable files CNVAuler bin and CNVRuler exe and one text file readme txt which is the change log of version history Simply uncompress it and select executable by type of your OS CNVRAuler bin for Linux CNVRuler exe for Windows Step 1 Download compressed zip file Right click and select Extract All fi C Adownload File Edit View Favorites Tools Help ack T a ya Search gt Folders Hub Address File and Folder Tasks Open mij Rename this file WHE search iy Move this File Explore Copy this File Extract All ES Publish this file to the Open With Web E mail this File gt Delete this File Cut Copy Send To Create Shortcut Delete ia PROGRAMS C Rename My Documents Cg Shared Documents ig My Computer Other
12. i Square Test Yates continuity correction Fisher s Exact Test 17 e Additional options for the association analysis 1 LRT CNV Ruler supports 2 Log Likelihood Ratio Test LRT and calculates p value of chi squared distribution of LRI With this value user can figure out the regression model used for association analysis is significantly better than null model or not Currently this option could be applied to logistic regression only 2 Population Stratification by PCA Since the association found could be due to the underlying structure of the population and not a disease associated locus CNV Ruler can use Principal Component Analysis PCA to adjust population stratification CNV Ruler calculates eigen vectors and uses up to 3 principal components as covariates for regression Currently this option could be applied to logistic regression only 3 Separated p values for Gain Loss lf a region contains both type of CNV Gain and Loss CNV Ruler will calculate p values for statistical test using only gain type CNVRs or only loss type ones with this option 4 Minor allele frequency Default value is 0 05 5 This means that CNVRs with less than 5 allele frequency will be excluded from the downstream association analysis Alternatively users can set their own threshold For example by setting the minor allele threshold to 0 users can observe the association result of all CNVRs regardless of the allele frequency D
13. l 0 0 H N W F AJOJO ARE z 00 w 0 v 5 y 3 3 426 602 4 132 545 132 780 234 793 142211 142578 11264 411 11 13 426 602 3 624 237 197 636 00 00 0 00 o 00 D D Eds e i Ss W BIS AU 0 S 4 in 242912 J O 5 5 D p D 2 19 References Bae JS et al 2010 Genome wide association analysis of copy number variations in subarachnoid aneurysmal hemorrhage J Hum Genet 55 11 726 30 Barnes C et al 2008 A robust statistical method for case control association testing with copy number variation Nat Genet 40 1245 1252 Colella S et al 2007 QuantiSNP an Objective Bayes Hidden Markov Model to detect and accurately map copy number variation using SNP genotyping data Nucleic Acids Res 35 2013 2025 Forer L et al 2010 CONAN copy number variation analysis software for genome wide association studies BMC Bioinformatic 11 318 Joeng Y et al 2008 CGHscape A Software Framework for the Detection and Visualization of Copy Number Alterations Genome amp Informatics 6 3 126 129 Korn J M et al 2008 Integrated genotype calling and association analysis of SNPs common copy number polymorphisms and rare CNVs Nat Genet 40 1253 1260 Pique Regi R et al 2010 R Gada a fast and flexible pipeline for copy number analysis in association studies BMC Bioinformatics 11 380 Purcell S et al 2007 P
14. n or loss information but have mean value of segmentation Therefore a cut off criterion is required to define the copy number gain or loss status Default value is 0 3 which means that a CNV segment with mean value lt 0 3 will be assigned as loss and gt 0 3 as gain CNV Users can set their own cut off filter 10 B Defining CNVR CNY Region Method Recurrence Gain Loss separated regions CNV Ruler supports three different definitions of CNV Regions CNVRs CNVR RO and Fragment They produce similar but slightly different boundaries and each of them has its own advantages and limitations as described in the main text e Method Select one of the following 3 definitions of CNVRs 1 CNVR CNV region 2 RO Reciprocal Overlap 3 Fragment El Samp le E le sample CNVRe Jhon 1 CNVR CNV region CNVR is defined by merging of overlapping CNVs CNVR trimming threshold Definition of CNVR is simple and straightforward but this definition can over estimate the size and frequency of CNVR due to the potential false calls which are usually rare and long sized CNV Ruler can trim these extreme ones during merging process by CNV frequencies In case of the CNVR method users can trim the sparse area by using the regional density recurrence threshold This option checks the regional density of participating CNVs base wise and trimming the sparse area not sa
15. of the association test You can sort it by any column by clicking its header It is also stored as a tab delimited text file The detailed option information is written in the header of the file example_ci_GC 20111217112307 x Chr Start End Size Freq Co Freq Ca Description pvalue _ 2 Log LRT LRT 68 683 835 69 942 276 1 258 442 2 Gain 0 220824 7 023 0 26 a 14 19 002 112 20 422 583 1 420 472 mixed 0 300951 7 863 0 3825 18 626 234 18 887 369 261 136 Gain 0 300951 7 863 0 3825 68 867 282 7 0 178 835 1 311 554 129 715 129 914 199 134 51 40 145 N N Loss 0 362467 8 179 0 4874 22 864 059 23 258 994 394 936 bee Gain 0 362467 8 179 0 4874 32 113 670 32 573 464 459 795 7 222 169 7 809 894 587 726 16 055 171 16 386 602 331 432 5 3 22 mixed 0 376290 8 227 0 5142 mixed 0 376290 8 227 0 5142 Gain 0 398321 0 4745 Gain Gain Gain 0 545384 8 mixed 0 585999 8 OJO E E o 00 N uu p Gain 0 362467 8 179 0 4874 88 861 135 89 182 355 321 221 00 00 00 00 I O N 00 Wl ojo uju SES 0 0 MIO 38 38 Z AJO 44 394 400 44 794 572 400 173 7 90 898 258 072 167 175 21 412 391 21 620 547 208 157 8 357 507 8 601 982 244 476 Gain 0 628756 816 Gain 0 628756 816 Gain 0 628756 8 816 Loss 0 687884 8 9 Gain 0 687884 i oss Gain 0 714802 8 929 Gain 0 746615 OJO SS a ou
16. se the possibility of false positive associations as well as the calculation burden NOTICE CNV region information is stored in imp directory and can be used later You can remove it safely by deleting tmp directory if error is occurred 14 e Examples of CNVR outputs from the same CNV data The list below contains the CNVs in chromosome 1 identified from the 7 samples from Affymetrix Genotype Console You can download the sample CNV file from our web site www ircgp com CNVRuler index html Total CNV List Chr Start End Type 1 61723 228694 Gain 1 61723 229063 Loss 1 61723 229607 Gain 1 61723 356530 Gain 1 85924 229607 Gain 1 235658 564621 Loss 1 740857 1030307 Gain 1 16830808 16935995 Gain 1 16968362 17298496 Gain 1 17029580 17245518 Gain 1 17035208 17177033 Gain 1 17036531 17182425 Gain 1 17037085 17182425 Gain 1 17045446 17190850 Gain 1 63704937 63810371 Gain 1 82461630 82644795 Gain 1 104130168 104307231 Gain 1 121343784 121482967 Gain 1 121343784 121482967 Gain 1 121343784 121482967 Gain 1 144036737 144849544 Gain 1 145206610 145398179 Gain 1 148530424 148662751 Gain 1 148530424 148953984 Gain 1 148947698 149051903 Loss 1 149086173 149202866 Loss 1 149086551 149190306 Loss 1 166574788 166966828 Gain 1 182454823 182611606 Loss 1 196706260 196812518 Gain 1 196706260 196812518 Gain 1 196711067 196812518 Gain 1 243163830 243274530 Gain 1 CNVR Recurrence Threshold is 0 1 Chr Start End Type 1 61 723 564 621 Mixed
17. tisfying the given density threshold default 0 1 This option does not affect RO nor Fragment method Additional options for building CNVRs Gain Loss separated region Using this option the CNVR can be created with same types of CNVs gain or loss type within the considering area If you select this option CNVR outputs will be copy number gain CNVR or loss CNVR If you don t select this option all overlapped CNVs will be used for building CNVR regardless of their type Ex Option is off Gain Loss option is on Gain c Loss Gain c Loss CNVR1 CNVR2 CNVR3 CNVR4 sample sample 2 Sample _ gt imple 1 E E E i CNVR CNVR CNVR GAIN 1 GAIN2 GAIN3 oss Region EA E CNVR CNVR CNVR LOSS 1 LOSS 2 LOSS 3 12 c E a 2 Reciprocal overlap RO CNV regions are determined by reciprocal overlap RO measure First CNVs which overlap at least one base are grouped as initial CNV clusters Within each cluster RO is calculated for each CNV to the others The pair of which RO is highest default minimum threshold is gt 50 will be merged and formed a CNV element in orange This process is repeated until every pair has RO of 50 or lower The detailed process is illustrated in the following figure CNV1 Initial CNV2 CNV QQ CNV3 Cluster 7 G VE CNV2 3 GF CNV4 CNV5 Y CNV1 Step 1 CNV1 CNV2 3 4 CNV5 Step 2
18. variables as many as you want Only the selected extra covariates will be included for the association analysis Example of clinical information file and data loaded screen Sample ID Phenotype Age Gender Smoking Atop ano 30 ss o af o aozas 18682 31 o o poe all ol 0 1 681 gt a SISIE 77 2 121 ial aa 6 Po k iso IM SI IO E arsi 199 75 op ol o Arez 2072 5 i of o 1 Analysis with sample ID and phenotype CNV Ruler ver 1 2 1beta Sample ID Sample ID Phenotype Phenotype h Covariate 2 Analysis with four main Cls CNV Ruler ver 1 2 1beta Clinical Information workspacelpaperiACAT1120108 example_ci manual Sample ID Gender eEnEWEE Age nn Covariate 3 Analysis with additional covariates CNV Ruler ver 1 2 1beta Clinical Information workspace papenACAT 20108 example_ci_manual bd Sample ID Sample ID Gender Gender w Phenotype Phenotype Age Covariate E Covariates Sample ID Phenotype Age Gender Smoking _ Atopy CNV Ruler ver 1 2 1beta Clinical Information workspace papenACAT 201 08 example_ci_manual tt Sample ID Sample ID Gender Gender Covariate i Phenotype Phenotype Age Age h e Step 2 Uploading CNV data Choose your CNV call output file for analyzing the CNVR based
Download Pdf Manuals
Related Search