migrate v20141106 from Google Code

ComMet/README

@@ -1,13 +1,21 @@
-== ComMet: an HMM-based approach to detection of differentially methylated regions
+== ComMet: an HMM-based approach to detection of differentially methylated regions (DMRs)
 
+version 1.1
 
-==== Required libraries 
 
-Boost C++ libraries 
-http://www.boost.org/
+==== Precompiled Binaries
+
+ComMet.v11.linux.64
+(Linux 3.2.0-4-amd64 + GCC 4.7.2 + Boost 1.49.0)
+
+ComMet.v11.linux.32
+(Linux 2.6.32-5-686 + GCC 4.4.5 + Boost 1.42.0)
 
 
-==== Build
+==== Build yourself
+
+Boost C++ libraries 1.35.0 or later are required.
+http://www.boost.org/
 
 $ tar xzvf ComMet-xxx.tar.gz
 $ cd ComMet-xxx
@@ -27,12 +35,6 @@ $ ComMet example/example.in example/example.out1 example/example.out2
 For more output DMRs,
 $ ComMet --threshold -5 example/example.in example/example.out1 example/example.out2
 
-For more accurate DMRs, 
-$ ComMet --dual example/example.in example/example.out1 example/example.out2
-
-Combining them, 
-$ ComMet --dual --threshold -5 example/example.in example/example.out1 example/example.out2
-
 
 ==== Input format
 
@@ -61,7 +63,7 @@ for plus and minus strands, and apply them to ComMet separately.
 
 ==== Output format
 
-output1 contains information of differential methylation at individual cytosines.
+output1 contains information of differential methylation at individual cytosine sites.
 See example/example.out1_
 
 Col.| Description
@@ -84,6 +86,30 @@ Col.| Description
 3   | 0-based genomic stop position
 4   | direction of differential methylation (UP/DOWN) comparing sample1 to sample2
 5   | log-likelihood ratio score 
+6   | log-likelihood ratio score divided by DMR length
+
+Make sure output1 and output2 are used properly considering the purpose of your study. 
+You should use output1 if you are interested only in differential methylation at 
+individual cytosine sites (Note that it is the purpose of most existing packages for 
+bisulfite sequencing data analysis developed by other groups).
+ComMet is mainly designed for DMR detection, i.e. determining precise boundaries of 
+regional differential methylation, even if DMRs include some cytosine sites whose 
+observed methylation changes are relatively weak due to limited sequencing depth. 
+Such an analysis is useful for identifying biologically important DMRs such as 
+cis regulatory elements; output2 is suitable for this purpose. 
+
+
+==== Tips for detection of DMRs in non-CpG context
+
+ComMet version 1.1 supports detection of DMRs in non-CPG context (CHG and CHH). 
+
+$ ComMet --noncpg example/example.chg.in example/example.chg.out1 example/example.chg.out2
+$ ComMet --noncpg example/example.chh.in example/example.chh.out1 example/example.chh.out2
+$ ComMet example/example.cpg.in example/example.cpg.out1 example/example.cpg.out2
+
+Make sure an input file is prepared separately for each context, and ComMet is executed 
+with proper options. We do not recommend that input files for different contexts are 
+combined, or ComMet is executed --noncpg option while an input file contains only CpGs.
 
 
 ==== FAQ
@@ -98,26 +124,32 @@ i.e. the 5'-CpG-3' in the plus strand, and the neighboring 3'-GpC-5' in the minu
 See the "Input format" section above for this issue. 
 Second, the input file may contain cytosines in non-CpG context; just remove them. 
 
-Q. Does ComMet support DMRs in non-CpG context (CHG or CHH)?
-A. 
-No. But we are planning to address this issue in the next version of ComMet.
-
 Q. The read counts in the example input file are decimals rather than integers. Why?
 A. 
 Either decimals or integers can be used for read counts in input files. 
 The reason that the example input file contains decimals is that some alignment tools produce 
 probability-weighted read counts. Of course, you can use your favorite aligners for preparing 
 input files that may contain integers only.
 
+Q. Can ComMet compute statistical significance (p-values) rather than likelihood ratio scores?
+A.
+No. But we are planning to address this issue in the next version of ComMet. 
+
 
 ==== History
 
-* version 1.0
+* version 1.1
+- implemented the algorithm described in [Saito et al., submitted, 2014]
+- implemented a testing version of algorithms for DMRs in non-CpG context 
+- added some tips in README about detection of DMRs in non-CPG context
 - tuned the default parameters
+
+* version 1.0
 - added the FAQ in README
+- tuned the default parameters
 
 * version 0.1 
-- implemented the algorithms described in [Saito et al., Nucleic Acids Res, 2013]
+- implemented the algorithms described in [Saito et al., Nucleic Acids Res, 2014]
 
 
 ==== Bisulfighter
@@ -138,7 +170,12 @@ http://creativecommons.org/licenses/by-nc-sa/3.0/
 
 Yutaka Saito, Junko Tsuji, and Toutai Mituyama,
 Bisulfighter: accurate detection of methylated cytosines and differentially methylated regions,
-Nucleic Acids Research, accepted for publication, 2013.
+Nucleic Acids Research, 42(6):e45, 2014. 
+
+Yutaka Saito, and Toutai Mituyama, 
+Detection of differentially methylated regions from bisulfite sequencing data 
+improved by hidden Markov models with new emission probabilities. 
+submitted. 
 
 
 ==== Contact