sh-miR Designer v2.0

Sh-miR designer is a project we started during iGEM 2013. This year we would like to continue and expand functionality of the software. sh-miR designer v1.0 (link) is aimed to create sh-miR molecules based on siRNAs provided by the user. In sh-miR designer v2.0 only the mRNA number (from NCBI database), which expression should be decreased can be provided. Moreover, we expanded functionality of the software with off-target validation and check of immune motifs and also extended miRNA-shuttles database. sh-miR Designer is a software aimed for fast and efficient design of effective RNA interference (RNAi) reagents - sh-miRs, also known as artificial miRNAs. sh-miRs are RNA particles whose structure is based on miRNA precursor pri-miRNA, but sequence interacting with transcript is changed depending on research purpose. Maintenance of structure of pri-miRNA is very important to enable cellular processing and therefore ensure functionality of artificial particles. sh-miRs delivered to cells on genetic vectors - plasmids or viral vectors - enter natural RNAi pathway and silence target mRNA. They can be used in genetic therapies and basic biomedical research. We will provide two applications to access the software, one which require siRNA sequences and the second which require transcript accession number from NCBI database. Each user will receive an account with login (e-mail) and password where he/she would be able to check history of software usage.
input form:

users' history:

Standard Parts

We are creating a particle which can be inserted into plasmid built by Standard parts from Registry of Standard Biological Parts.

User Guide

Usage via Web server

go to our website shmir.wmi.amu.edu.pl

Usage via client

Fold sequence via mfold:
./shmir_client.py mfold ACTGAUUUGAC
Create sh-miR from one siRNA strand (active) or two siRNa strands separated by space.
First strand is active, both are in 5-3 orientation

./shmir_client.py from_sirna UUUGUAUUCGCCCUAGCGC CGCUAUGGCGAAUACAAACA
Create sh-miR from transcript name.
Optional parameters
--min_gc : Minimal "GC" content in strand -- default: 40
--max_gc : Maximal "GC" content in strand -- default: 60
--max_offtarget : Maximal offtarget in strand -- default: 10
--mirna_name : The name of miRNA backbone to use -- default: "all"
--stymulators : One of ["yes", "no", "no_difference"] -- default: "no_difference"
./shmir_client.py from_transcript NM_001618.3 --mirna_name miR-30a --min_gc 22

Technology

The application was built with Python 2.7. It is divided into 3 parts:
RESTful API - the “heart” of our software - it performs all calculations. Our application is built on CentOS (Linux) operating system. To maintain better compatibility with our server we use Vagrant to create Virtual Machine and Puppet to manage infrastructure of our project. All you need to do is: type into console “vagrant up” and after a while everything is ready to work. This subproject is written with following technologies/frameworks/libraries: Flask, Flask-Cache, SqlAlchemy, PostgreSQL, Celery, Flower, Rabbitmq, Redis, BioPython, mfold (lightweight), uwsgi, sphinx, Python testing tools (mock, nose, unittest).
Console client - simple console application which communicates with API using requests library. It has 3 options: mfold (folding structure), from_sirna (create sh-miR from siRNA) and from_transcript (create sh-miR from transcript name).
Website - application written in Django framework. It enables you to use the API via graphical interface. You can also create your own account and save the history of your requests for later usage.

API documentation

Here you can find api documentation from code (docstrings): link unavailable yet

API for developers

Here you can find api for developers: https://github.com/Nozdi/shmir

Releases (re-use of 2013iGEM software)

sh-miR designer v1.0
First version of the program presented on iGEM 2013 by our team. The user have to provide siRNA sequence (one or both strands). The miRNA database includes 5 miRNAs.
sh-miR designer v1.1
Strand discrimination functionality was added based on miRbase.
extended miRNA database
sh-miR designer v2.0
siRNA prediction algorithm was added to functionality (the input is NCBI transcript number)
user can choose GC content of siRNA molecule, maximal off-target, if he/she wants immunostimulatory sequences inside the molecule and to choose if he/she wants to include all or chosen scaffold
extended miRNA database
What we are working on:
strand discrimination based on therodynamics of siRNA ends;
preparing construct with flanking restriction enzymes
providing plasmids with parts from standard parts

Unit Tests

A few words about

In our application we are using unit tests to test alghoritms, functions and database.
We are testing inputs and outputs:
def test_validate_gc_content(self):
is_in_range = validate_gc_content('ACGT', 40, 60)
self.assertTrue(is_in_range)
Caught errors:
def multiple_input(self, data, error):
for strand in data:
with self.assertRaises(errors.InputException) as err:
validators.check_input(strand)
self.assertEqual(error, str(err.exception))
Database models:
def test_get_and_put_to_db(self):
new_obj = self.TestModel(id=153)
self.put_to_db(new_obj)
objs = self.get_all(self.TestModel)
self.assertEqual(len(objs), 1)

We are using Mock library to test our functions which are using external API’s to create fake response from server and return them into our function:
@patch.object(ncbi_api.Entrez, 'esearch')
@patch.object(ncbi_api.Entrez, 'efetch')
def test_get_data(self, mock_efetch, mock_esearch):
ids = [69]
mock_read = Mock()
mock_read.read.return_value = json.dumps({'esearchresult': {'idlist': ids}})
mock_esearch.return_value = mock_read
ncbi_api.get_data(self.GOOD_MRNA)
mock_esearch.assert_called_with(
term='NM_001128164.1', db=self.DATA_BASE, retmode='json'
)
mock_efetch.assert_called_once_with(
db=self.DATA_BASE, id=ids, rettype='fasta', retmode='text'
)

Validation

old version shmiR v1.0 (average time)

real 0m6.854s
user 0m0.063s
sys 0m0.027s

new version shmiR v1.1 (average times)

real 0m3.271s
user 0m0.155s
sys 0m0.041s

Test cases

1) ATXN3 - therapeutic sh-miR against SCA3 disease
transcript number: NM_001164782.1
GC content: default
Off-target: 10
scaffold: ALL
immunostimulatory: NO
2) P54 - therapeutic sh-miR against cancer
transcript number: NM_000546.5
GC content: default
Off-target: 10
scaffold: ALL
immunostimulatory: YES
3) biological function analysis, PARP1, we have miR-30a in our lab
transcript number: NM_001618.3
GC content: default
Off-target: 0
scaffold: miR-30a
immunostimulatory: no_difference (default)
4) biological function, PARP1, we have miR-30a in our laboratory and have immune cells
transcript number: NM_001618.3
GC content: default
Off-target: 0
scaffold: miR-30a
immunostimulatory: NO
5) biological function, PARP1 and have immune cells
transcript number: NM_001618.3
GC content: default
Off-target: 0
scaffold: ALL
immunostimulatory: NO

References

Judging Form (in bold what we did to achieve gold medal)