Recent changes to Home

Home modified by Junho Kim

Junho Kim — Tue, 01 Jan 2019 11:31:25 -0000

--- v25
+++ v26
@@ -53,6 +53,7 @@
     git clone https://junhokim@git.code.sf.net/p/replow/code replow-code

     # build RePlow
+    cd replow-code
     mvn package    

     # test running. You should correct the path for REF and RSCRIPT in run.sh

Home modified by Junho Kim

Junho Kim — Tue, 01 Jan 2019 09:55:38 -0000

--- v24
+++ v25
@@ -50,7 +50,7 @@
     cd RePlow

     # get the source code
-    git clone http://
+    git clone https://junhokim@git.code.sf.net/p/replow/code replow-code

     # build RePlow
     mvn package

Home modified by Junho Kim

Junho Kim — Tue, 01 Jan 2019 09:53:06 -0000

--- v23
+++ v24
@@ -16,6 +16,7 @@

 * [Rscript](https://stat.ethz.ch/R-manual/R-devel/library/utils/html/Rscript.html)
 * [fitdistrplus](https://cran.r-project.org/web/packages/fitdistrplus/fitdistrplus.pdf)
+
 


 Installation of RePlow
@@ -41,6 +42,23 @@
     : Predefined high mapQ regions for the estimation of background error rate. For computational efficiency, only shared regions between high mapQ and given target regions are used for the error rate estimation if the total size of target region > 1Mb. 
 *  **RePlow.license**
     : Software licensing terms
+
+Or you can directly build RePlow from the source code by yourself. To compile RePlow, Java JDK 8 and Maven 3.0+ are required. You'll get RePlow-1.x.y.jar in your folder if the source code is successfully compiled.
+
+    # make a new source directory
+    mkdir RePlow
+    cd RePlow
+    
+    # get the source code
+    git clone http://
+    
+    # build RePlow
+    mvn package    
+    
+    # test running. You should correct the path for REF and RSCRIPT in run.sh
+    cd testdata
+    ./run.sh
+
 


 Running RePlow
@@ -112,12 +130,12 @@
 Test running
 =======

-We have provided a small demo dataset for installation verification. You can find the data and do test run under testdata folder.
+We have provided a small demo dataset for installation verification. You can find the data and do test run under testdata folder. Before running this, you should change the path for REF and RSCRIPT in run.sh file.

     cd testdata
     ./run.sh

-or you can directly test the running command
+You can also test the running command directly.

     java -jar ../RePlow.jar \
    -r Reference.fasta \
@@ -219,9 +237,9 @@
 Single running mode in RePlow
 =======

-Although RePlow is designed to simultaneously analyze multiple replicates, it can also be used for a single replicate. On-the-fly estimation of background errors provides great benefits to remove false positives even for the analysis of a single library (See the figure below). If you are trying to detect low-VAF mutations with deep-enough data and you don't have any replication for this, you can try our single running mode to eliminate false positive error calls. However, when the sequencing data contains a large number of background errors with VAFs similar to true mutations (e.g. ILA cases in the figure), the VAF cutoff becomes high enough to also remove true mutations due to the high estimated error rate, resulting in significant decrease of sensitivity. With replicates of those error-rich data, RePlow can recover the removed true mutations based on the VAF concordance between replicates therefore showed far higher sensitivity. In addition, RePlow with replicates provides more sophisticated filtration of false positives, considering the error profile concordance; false positives with exceptionally high VAFs (therefore beyond the cutoff) from the analysis of single library can be subsequently filtered out when multiple replicates are considered together, based on their concordance of VAF and error probability. So all modules including the ones work well for the single library have benefits of replication, thus we recommend to generate library replicates for accurate detection of low-level mutations.
+Although RePlow is designed to simultaneously analyze multiple replicates, it can also be used for a single replicate. On-the-fly estimation of background errors provides great benefits to remove false positives even for the analysis of a single library (See the figure below). If you are trying to detect low-VAF mutations with deep-enough data and you don't have any replication for this, you can try our single running mode to eliminate false positive error calls. However, when the sequencing data contains a large number of background errors with VAFs similar to true mutations (e.g. ILA cases in the figure), the VAF cutoff becomes high enough to also remove true mutations due to the high estimated error rate, resulting in significant decrease of sensitivity. With replicates of those error-rich data, RePlow can recover the removed true mutations based on the VAF concordance between replicates therefore showed far higher sensitivity. In addition, RePlow with replicates provides more sophisticated filtration of false positives, considering the error profile concordance; false positives with exceptionally high VAFs (therefore beyond the cutoff) from the analysis of single library can be subsequently filtered out when multiple replicates are considered together, based on their concordance of VAF and error probability. So all modules including the ones work well for the single library have benefits of replication, thus we recommend to generate library replicates or just use single running mode to prioritize high confidence candidates for accurate detection of low-level mutations.

 
-
+

Home modified by Junho Kim

Junho Kim — Mon, 31 Dec 2018 23:15:55 -0000

--- v22
+++ v23
@@ -221,5 +221,7 @@

 Although RePlow is designed to simultaneously analyze multiple replicates, it can also be used for a single replicate. On-the-fly estimation of background errors provides great benefits to remove false positives even for the analysis of a single library (See the figure below). If you are trying to detect low-VAF mutations with deep-enough data and you don't have any replication for this, you can try our single running mode to eliminate false positive error calls. However, when the sequencing data contains a large number of background errors with VAFs similar to true mutations (e.g. ILA cases in the figure), the VAF cutoff becomes high enough to also remove true mutations due to the high estimated error rate, resulting in significant decrease of sensitivity. With replicates of those error-rich data, RePlow can recover the removed true mutations based on the VAF concordance between replicates therefore showed far higher sensitivity. In addition, RePlow with replicates provides more sophisticated filtration of false positives, considering the error profile concordance; false positives with exceptionally high VAFs (therefore beyond the cutoff) from the analysis of single library can be subsequently filtered out when multiple replicates are considered together, based on their concordance of VAF and error probability. So all modules including the ones work well for the single library have benefits of replication, thus we recommend to generate library replicates for accurate detection of low-level mutations.

-[[img src=Sup_fig14.png alt=Performance assessment of RePlow with the single and duplicate mode]]
-

+
+
+
+

Home modified by Junho Kim

Junho Kim — Mon, 31 Dec 2018 22:47:55 -0000

--- v21
+++ v22
@@ -221,6 +221,5 @@

 Although RePlow is designed to simultaneously analyze multiple replicates, it can also be used for a single replicate. On-the-fly estimation of background errors provides great benefits to remove false positives even for the analysis of a single library (See the figure below). If you are trying to detect low-VAF mutations with deep-enough data and you don't have any replication for this, you can try our single running mode to eliminate false positive error calls. However, when the sequencing data contains a large number of background errors with VAFs similar to true mutations (e.g. ILA cases in the figure), the VAF cutoff becomes high enough to also remove true mutations due to the high estimated error rate, resulting in significant decrease of sensitivity. With replicates of those error-rich data, RePlow can recover the removed true mutations based on the VAF concordance between replicates therefore showed far higher sensitivity. In addition, RePlow with replicates provides more sophisticated filtration of false positives, considering the error profile concordance; false positives with exceptionally high VAFs (therefore beyond the cutoff) from the analysis of single library can be subsequently filtered out when multiple replicates are considered together, based on their concordance of VAF and error probability. So all modules including the ones work well for the single library have benefits of replication, thus we recommend to generate library replicates for accurate detection of low-level mutations.

-![Performance assessment of RePlow with the single and duplicate mode](https://sourceforge.net/p/replow/wiki/Home/attachment/Sup_fig14.png)
-
-

+[[img src=Sup_fig14.png alt=Performance assessment of RePlow with the single and duplicate mode]]
+

Home modified by Junho Kim

Junho Kim — Mon, 31 Dec 2018 22:45:06 -0000

--- v20
+++ v21
@@ -221,6 +221,6 @@

 Although RePlow is designed to simultaneously analyze multiple replicates, it can also be used for a single replicate. On-the-fly estimation of background errors provides great benefits to remove false positives even for the analysis of a single library (See the figure below). If you are trying to detect low-VAF mutations with deep-enough data and you don't have any replication for this, you can try our single running mode to eliminate false positive error calls. However, when the sequencing data contains a large number of background errors with VAFs similar to true mutations (e.g. ILA cases in the figure), the VAF cutoff becomes high enough to also remove true mutations due to the high estimated error rate, resulting in significant decrease of sensitivity. With replicates of those error-rich data, RePlow can recover the removed true mutations based on the VAF concordance between replicates therefore showed far higher sensitivity. In addition, RePlow with replicates provides more sophisticated filtration of false positives, considering the error profile concordance; false positives with exceptionally high VAFs (therefore beyond the cutoff) from the analysis of single library can be subsequently filtered out when multiple replicates are considered together, based on their concordance of VAF and error probability. So all modules including the ones work well for the single library have benefits of replication, thus we recommend to generate library replicates for accurate detection of low-level mutations.

-![Performance assessment of RePlow with the single and duplicate mode](http://!))
-
-

+![Performance assessment of RePlow with the single and duplicate mode](https://sourceforge.net/p/replow/wiki/Home/attachment/Sup_fig14.png)
+
+

Home modified by Junho Kim

Junho Kim — Mon, 31 Dec 2018 22:43:18 -0000

Home modified by Junho Kim

Junho Kim — Mon, 31 Dec 2018 21:57:02 -0000

--- v18
+++ v19
@@ -130,6 +130,14 @@

 You will get an output file (out.snv.call) under OUTPUT folder  when your running is successfully finished.

+
+

+
+Resource requirements
+=======
+
+To analyze a 500X whole-exome sequencing dataset (~25 Gb), RePlow takes ~9 hours per sample for the entire process with a single core and ~6.5 Gb RAM usage. 
+RePlow is not suitable for whole-genome datasets since it is designed to focus on detecting low-VAF mutations (~1% VAF), which requires deep enough depth for sequencing data, like 400X at least.

 


@@ -205,3 +213,14 @@
     : List of the variant candidates with the probability scores. All rejected candidates are also included with their filtered reasons. Final somatic candidates are tagged as 'PASS' in the Filter field. Labeled name (with -L) or the name of the first BAM file from the given input list will be used as an ID for the output files.
 +  **intersection.bed (optional)**
     : If the target regions are different between given replicates, an intersection is calculated first and saved as a file (intersection.bed). Only the regions included in the intersection file is considered for the entire analysis.
+
+

+
+Single running mode in RePlow
+=======
+
+Although RePlow is designed to simultaneously analyze multiple replicates, it can also be used for a single replicate. On-the-fly estimation of background errors provides great benefits to remove false positives even for the analysis of a single library (See the figure below). If you are trying to detect low-VAF mutations with deep-enough data and you don't have any replication for this, you can try our single running mode to eliminate false positive error calls. However, when the sequencing data contains a large number of background errors with VAFs similar to true mutations (e.g. ILA cases in the figure), the VAF cutoff becomes high enough to also remove true mutations due to the high estimated error rate, resulting in significant decrease of sensitivity. With replicates of those error-rich data, RePlow can recover the removed true mutations based on the VAF concordance between replicates therefore showed far higher sensitivity. In addition, RePlow with replicates provides more sophisticated filtration of false positives, considering the error profile concordance; false positives with exceptionally high VAFs (therefore beyond the cutoff) from the analysis of single library can be subsequently filtered out when multiple replicates are considered together, based on their concordance of VAF and error probability. So all modules including the ones work well for the single library have benefits of replication, thus we recommend to generate library replicates for accurate detection of low-level mutations.
+
+![Performance assessment of RePlow with the single and duplicate mode](http://!))
+
+

Home modified by Junho Kim

Junho Kim — Mon, 31 Dec 2018 19:55:28 -0000

--- v17
+++ v18
@@ -114,14 +114,11 @@

 We have provided a small demo dataset for installation verification. You can find the data and do test run under testdata folder.

--------------------------------------
     cd testdata
     ./run.sh
--------------------------------------

 or you can directly test the running command

--------------------------------------
     java -jar ../RePlow.jar \
    -r Reference.fasta \
    -b rep1.bam,rep2.bam \
@@ -130,7 +127,6 @@
    -R RSCRIPT_PATH \
    -o OUTPUT_DIR \
    -L LABEL_FOR_OUTPUT
--------------------------------------

 You will get an output file (out.snv.call) under OUTPUT folder  when your running is successfully finished.

Home modified by Junho Kim

Junho Kim — Mon, 31 Dec 2018 19:54:42 -0000

--- v16
+++ v17
@@ -106,6 +106,34 @@
     -k, --keep_intermediate_files               Keep intermediate files

 -------------------------------------------
+
+

+
+Test running
+=======
+
+We have provided a small demo dataset for installation verification. You can find the data and do test run under testdata folder.
+
+-------------------------------------
+    cd testdata
+    ./run.sh
+-------------------------------------
+
+or you can directly test the running command
+
+-------------------------------------
+    java -jar ../RePlow.jar \
+   -r Reference.fasta \
+   -b rep1.bam,rep2.bam \
+   -N normal.bam \
+   -T test.bed \
+   -R RSCRIPT_PATH \
+   -o OUTPUT_DIR \
+   -L LABEL_FOR_OUTPUT
+-------------------------------------
+
+You will get an output file (out.snv.call) under OUTPUT folder  when your running is successfully finished.
+