diff --git a/README b/README
index b4b517148f3fe361e85adf1d144fe7f78b2bb556..82a0d7ae0ee29387b1c01fd7b4503eae105ffa4d 100644
--- a/README
+++ b/README
@@ -19,6 +19,15 @@ cd pirs
 make
 
 
+################
+# Requirements #
+################
+
+- Python 2
+- Python 3
+- BioPython for Python3
+
+
 ################
 # How to use ? #
 ################
@@ -61,4 +70,21 @@ We use SVsim to generate positions of SVs. https://github.com/GregoryFaust/SVsim
 
 In the output directory there are:
 - A vcf file (genotypes.vcf) that list for each SV their position and the genotype for each individual
-- For each individual, 2 fastq files of the paired end reads (INDIV_%d_100_180_(1|2).fq.gz)
\ No newline at end of file
+- For each individual, 2 fastq files of the paired end reads (INDIV_%d_100_180_(1|2).fq.gz)
+
+
+###############
+# Limitations #
+###############
+
+As pirs works only on Intel CPUs, this software is only compatible with Intel. Tested and failed on AMD CPUs.
+
+
+######################
+# External softwares #
+######################
+
+This software use several external programs:
+- SVsim (included): https://github.com/GregoryFaust/SVsim
+- pirs (included): https://github.com/galaxy001/pirs
+- SeqIO from BioPython: http://biopython.org/wiki/SeqIO
\ No newline at end of file