Updated release notes

docs/release_notes_0.3.2.md

 # Release notes Hotfix 0.3.2
 
-* There was a bug found in our RNA seq pipeline ( Gentrap )
+* A bug was discovered in our RNA seq pipeline ( Gentrap )
  * The merged count table missed out on 1 gene consistently. The separate count files per sample where not affected by this bug
 * This bug involved all previous runs from Gentrap of the last 2 months, all customers where informed and if requested a new file is delivered.
- * we manually checked all the runs and found that the first gene was almost never expressed in our data sets
+ * We manually checked all the runs and found that the first gene was almost never expressed in our data sets
+ * In case you haven't heard from us, you can assume your data not to have been affected
  
\ No newline at end of file

docs/release_notes_0.4.0.md

 # Release notes Biopet version 0.4.0
 
-Since our last release a couple of things are added, changed or fixed:
-
-* A reporting framework is added for most pipelines
- * This framework produces a static HTML report which can be viewed either offline or online
+* A reporting framework has been added for most pipelines
+ * This framework produces a static HTML report which can be viewed in your browser
  * The framework contains lots of quality control and downstream analyses plots (genome coverage, transcript coverage etc etc.)
-* A problem with an Obscure `NullPointerException` this was thrown if `output_dir` was not set in config. This now gives a nice error message which points to the missing key in the config
-* Pipelines now automatically right a config file if none is specified on commandline
-* VariantContextWriterBuilder does not fail anymore when the input is not a block_compressed_VCF
+* An issue where a NullPointerException was being thrown when output_dir was not set in the config was fixed. This now gives a nice error message which points to the missing key in the config
+* Pipelines now automatically write a log file if none is specified on command line
+* Tools writing to VCF will no longer fail when the *output* is not a *gzipped* VCF
 * Pipelines now support passing config options directly into the commandline prompt
 * Pipelines now support a more readable config file format [YAML](https://en.wikipedia.org/?title=YAML)
-* Memlimit and vmem memory issues are solved by increasing the amount of available memory when a job fails. ```--retry 5``` should do the trick
-* Samtools Mpileup option ```-l``` is changed from optional to input
-* BamMetrics is now capable of working with the reporting framework
+* Memlimit and vmem memory issues are solved by automatically increasing the amount of available memory when a job fails. ```--retry 5``` should do the trick
 * BamMetrics pipeline is updated to work with newest version of Picard
-* VCF stats has a stricter sample to sample overlap, since in the previous versions these settings were to loose
-* VCF stats is now capable of summarizing stats per bin (bin size is changeable)
+* A bug in VcfStats in comparing samples alleles is fixed. Now each allele can only be used once in the comparison.
+* VcfStats is now capable of summarizing stats per bin (bin size is changeable)
 * There is now a module which checks for the present of correct reference files if not it automatically builds the appropriate ref files
 
-
-
-Some pipelines where updated as well:
+Some pipelines were updated as well:
 
 * Gentrap
+* Shiva
\ No newline at end of file