diff --git a/README.md b/README.md
index 437fe1a3a5688eb63ac491428a160497ae1fb41c..46a7a606629260e0c836e875dde2e87bdcae45e6 100644
--- a/README.md
+++ b/README.md
@@ -2,7 +2,7 @@
 The software was intentionally designed to separate the data input and analysis tools.
 
 Similar to JUNO Offline, two main structures exist: one is named `**algorithm**`, and the other is named `**tool**`. Each of these structures has its own set of rules for definition.
-1. `**algorithm**` can accept the input **arguments** and input **data** (only `ROOT` data can be accepted now). The **data** is input as a list written in a text file, **arguments** can be delivered to the `algorithm` by setting in the python script. Besides, the `**algorithm**` can call analysis `tool` to process the data.
+1. **`algorithm`** can accept the input **arguments** and input **data** (only `ROOT` data can be accepted now). The **data** is input as a list written in a text file, **arguments** can be delivered to the `algorithm` by setting in the python script. Besides, the `**algorithm**` can call analysis `tool` to process the data.
 2. `**tool**` contains the true analysis algorithm
 
 ## Create your own algorithm
@@ -70,7 +70,7 @@ More detail can be found in `./Analysis/CommissioningAlg/`.
 
 ## Create your own analsis tool
 
-Similar to `ExampleAlg`, the folders, CMakeLists.txt and __init__.py should be created
+Similar to `ExampleAlg`, the folders, CMakeLists.txt and \_\_init\_\_.py should be created
 
 A `**tool**` is defined as a `c++ class` too.