diff --git a/GettingStarted.md b/GettingStarted.md
index 9900aa7..99e26ac 100644
--- a/GettingStarted.md
+++ b/GettingStarted.md
@@ -28,8 +28,8 @@ in _pytor_ file.
 
 CNVpytor will detect reference genome and use internal database for GC content and 1000 genome strict mask.
 
-This works for hg19 and hg38 genomes. For other species or reference genomes you have to 
-[specify reference genome](examples/AddReferenceGenome.md).
+After instalation this works for hg19 and hg38 genomes. For other species or reference genomes you have to 
+[describe reference genome](examples/AddReferenceGenome.md).
 
 To check is reference genome detected use:
 
@@ -46,6 +46,7 @@ Using reference genome: hg19 [ GC: yes, mask: yes ]
 
 
 
+
 First hose bin size. It has to be divisible by 100. Here we will use 10 kbp and 100 kbp bins.
 
 To calculate binned, GC corrected RD signal type:
diff --git a/cnvpytor/genome.py b/cnvpytor/genome.py
index b16ac70..d59fcfc 100644
--- a/cnvpytor/genome.py
+++ b/cnvpytor/genome.py
@@ -282,8 +282,7 @@ def load_reference_genomes(cls, filename):
 
         """
         _logger.info("Reading configuration file '%s'." % filename)
-        import_reference_genomes = {}
-        exec(open(filename).read())
+        exec(open(filename).read(),globals())
         for g in import_reference_genomes:
             _logger.info("Importing reference genome data: '%s'." % g)
             cls.reference_genomes[g] = import_reference_genomes[g]
diff --git a/examples/AddReferenceGenome.md b/examples/AddReferenceGenome.md
new file mode 100644
index 0000000..8123ff1
--- /dev/null
+++ b/examples/AddReferenceGenome.md
@@ -0,0 +1,79 @@
+# Configuring reference genome
+
+For GC correction and 1000 genome strict mask filtering CNVpytor uses information 
+related to the reference genome. With installation two reference genomes are
+available: hg19 (GRCh37) and hg28 (GRCh38).
+
+If you want to use other reference genome for human or other species first we have 
+to create GC and mask file (optional).
+
+In this example we will configure mouse reference genome MGSCv37.
+
+To create GC file we need sequence of the reference genome in fasta.gz file:
+
+```
+> cnvpytor -root MGSCv37_gc_file.pytor -gc ~/hg19/mouse.fasta.gz -make_gc_file
+```
+
+This command will produce _MGSCv37_gc_file.pytor_ file that contains information about 
+GC content in 100-base-pair bins.
+
+For reference genomes where we have strict mask in the same format as 100 Genomes Project 
+[strict mask](http://ftp.1000genomes.ebi.ac.uk/vol1/ftp/data_collections/1000_genomes_project/working/20160622_genome_mask_GRCh38/),
+we can create mask file using command:
+
+```
+> cnvpytor -root MGSCv37_mask_file.pytor -mask ~/hg19/mouse.strict_mask.whole_genome.fasta.gz -make_mask_file
+```
+
+If we do not have mask file, we can skip this step. Mask file contains information about 
+regions of the genome that are more accessible to next generation sequencing methods 
+using short reads. CNVpytor uses P marked positions to filter SNP-s and read depth signal.
+If reference genome configuration does not contain mask file, CNVpytor will still be fully functional, 
+apart from the filtering step.
+You may also generate your own mask file by creating fasta file that contains character "P" if corresponding 
+base pair passes the filter and any character different than "P" if not.
+
+Now, we will create example_ref_genome_conf.py file containing following:
+
+```
+import_reference_genomes = {
+    "mm9": {
+        "name": "MGSCv37",
+        "species": "Mus musculus",
+        "chromosomes": OrderedDict(
+            [("chr1", (197195432, "A")), ("chr2", (181748087, "A")), ("chr3", (159599783, "A")),
+            ("chr4", (155630120, "A")), ("chr5", (152537259, "A")), ("chr6", (149517037, "A")),
+            ("chr7", (152524553, "A")), ("chr8", (131738871, "A")), ("chr9", (124076172, "A")),
+            ("chr10", (129993255, "A")), ("chr11", (121843856, "A")), ("chr12", (121257530, "A")),
+            ("chr13", (120284312, "A")), ("chr14", (125194864, "A")), ("chr15", (103494974, "A")),
+            ("chr16", (98319150, "A")), ("chr17", (95272651, "A")), ("chr18", (90772031, "A")),
+            ("chr19", (61342430, "A")), ("chrX", (166650296, "S")), ("chrY", (15902555, "S")),
+            ("chrM", (16299, "M"))]),
+        "gc_file":"/..PATH../MGSCv37_gc_file.pytor",
+        "mask_file": "/..PATH../MGSCv37_mask_file.pytor"
+    }
+}
+```
+
+Last line can be skipped, if there is no mask file. 
+
+To use CNVpytor with new reference genome us -conf option in each cnvpytor command, e.g.
+```
+cnvpytor -conf REL_PATH/example_ref_genome_conf.py -root file.pytor -rd file.bam
+```
+
+CNVpytor will use chromosome lengths from alignment file to detect reference genome. 
+However, if you configured reference genome after you had already run -rd step you 
+could assign reference genome using -rg:
+```
+cnvpytor -conf REL_PATH/example_ref_genome_conf.py -root file.pytor -rg mm9
+```
+
+To avoid typing "-conf REL_PATH/example_ref_genome_conf.py" each time you run cnvpytor, 
+you can create an alias. However, we would like to encourage you to send us configuration, 
+gc and mask file and we would be glad to include it into the CNVpytor code. Or, even better, 
+fork the repository on GitHub, add configuration in cnvpytor/genome.py, data files in cnvpytor/data 
+and create a pull request.
+
+