LX-miR Design

This repository contains all the code used to create the miRNA-targeting LX-miR CRISPR library. To keep the relationship between primary miRNAs, mature miRNAs, and sgRNAs clear, a relational (MySQL or MSSQL) database is created.

Dependencies

• Python 2.7
• Jupyter - notebooks for organizing code
• data_processing - helper package with database connect and editing functions
• MySQL.Connector - connects to mySQL database
• Pandas - dataframes
• Paramiko - SSH connection
• PyODBC - connects to MSSQL database
• RNAfold - RNA secondary structure prediction
• SciPy - scientific computing functions
• SSHTunnel - SSH tunnel

Database Summary

The code in this repository will create a database with 6 tables. Below is a summary of this database.

The PrimaryMicroRNA table contains information about all 1881 human primary microRNAs. The PriID is the accession number for the miRNA and is the primary key for this table. The chromosome on which the primary miRNA is located is stored as 1-22 or X or Y. The StemloopSeq contains the DNA sequence of the stemloop as anotated in miRBase. The LongSeq is the stemloop sequence plus an additional 20 bp on either side, which allows for all sgRNA with cutting sites in the stemloop to be identified. The RNAfold dot bracket structure of the stemloop sequence is also provided, which helps determine where in the stemloop the sgRNA will cleave. The miRNA family name is provided if the primary miRNA is known to be part of a miRNA family. The HighConfidence variable is a 'T' if the primary miRNA is considered high confidence (see this blog post and this paper for more about the high confidence anotation).

The MatureMicroRNA table contains information about the mature microRNAs. The MatID is the accestion number for the mature miRNA and is the primary key for this table. The PriID corresponds to the PriID in the PrimaryMiRNA table and has a foreign key constraint.

The SingleGuideRNA table contains information about each unique sgRNA. This table does not contain any information about where the sgRNA targets due to the fact some sgRNAs target multiple miRNAs. The SgID is an automatically generated number will no meaning. The exclude field contains a single character describing why the sgRNA should be excluded from the pool:

• 'R': Greater than 10 targets in the hg19 genome with no mismatches
• 'T': 4 or greater T's in a row, which can lead to RNA Pol III termination
• 'A': All targets of the sgRNA have an NAG PAM, which is much less effective than an NGG PAM
• 'C': The cleavage site is outside of the stemloop
• 'D': Less than 10, but at least one off-target site which exactly matches the sgRNA
• 'L': Low max Zhang score (<0.2)

If multiple possible exclussion reasons apply, the reason most likely to affect the sgRNA read out was used ('R'>'T'>'A'>'C'>'D'>'L'). The presence of the exact sgRNA sequence in the Zhang lentiCRISPRv2 pool was recorded as a 'Y' in the ZhangLibrary column. The scores for each sgRNA using various scoring methods are also included in this table. The Zhang lab sgRNA scoring website was scripted against (see Zhang Scoring) and the scores were recorded in the ZhangScore column. The CRISPRseek scores for the first 1,000 sgRNAs are included, though running this R package was too resource intensive to perform for all sgRNAs. On-target scores for each sgRNA were provided by sgRNAScorer (see this paper), the orgional Doench scoring algorthm(see this paper) and an updated version of Doech, Azimuth (described in this bioRXiv paper and this Nature Biotechnology paper). The number of exact matches to the sgRNA sequence in the hg19 genome, as found by alignment using Bowtie2, are also recorded. The sequence of the oligo to be ordered (85 nts long), which includes primer binding sites, is included in the table. The final element in the table is a boolean value stating if the give sgRNA is or is not included in the ordered miR-ome oligos.

The SgRNATargetInformation table contains information about each miRNA site targeted by the sgRNAs. This table contains any data specific to a single site, not to the sgRNA which can target multiple sgRNAs. This includes the LongSg sequence: NNNN[sgRNA target 20]NGGNNN, which is used for Doench/Azimuth scoring (see above). The PriID of the targeted miRNA corresponds to the PriID in the PrimaryMiRNA table and has a foreign key constraint. If the cleavage site falls within a mature miRNA site, the MatID also corresponds the the MatID in the MatureMiRNA table and has a foreign key constraint. The SgID also has a foreign key constraint with the SgIDs in the SingleGuideRNA table.

The OverlappingSgRNAs table contains information about which sgRNAs have identical seeds. These sgRNAs which target the same miRNA are likely to have similar off-target sites and may create false positives. Therefore, only one sgRNA with any given seed is used when there are enough sgRNA to allow for the top four to be choosen. The SelectedSgID has a higher ZhangScore than the OverlappingSgID, which is discarded.

The ControlSgRNA table contains the name and the sequence of the control sgRNAs used along with the ID originally used in the paper from whtich the control sgRNA was taken.

The InPool table contains information the sgRNAs which are included in the miRNA-ome pool. Includes the id of the sgRNA or the name of the sgRNA (for controls) and the sequence of the ssDNA oligo submitted for plate synthesis.

PrimaryMicroRNA

Variable	Type
PriID	varchar(50)
PriMiRName	varchar(50)
Chr	varchar(2)
ChrStrand	char(1)
GenomeStart	int
GenomeEnd	int
StemLoopSeq	varchar(200)
LongSeq	varchar(250)
RNAfold	varchar(200)
MiRFamily	varchar(50)
HighConfidence	nchar(1)

MatureMicroRNA

Variable	Type
MatID	varchar(50)
MatMiRName	varchar(50)
PriID	varchar(50)
MatStart	int
MatEnd	int
MatSeq	varchar(50)

SingleGuideRNA

Variable	Type
SgID	int
SgRNA	char(20)
Exclude	char(1)
ZhangLibrary	varchar(1)
ZhangScore	float
CRISPRseek	float
SgRNAScorer	float
MaxDoenchScore	float
MaxAzimuthScore	float
NumExactMatch	int
hg38NumExactMatch	int

SgRNATargetInformation

Variable	Type
SgID	int
LongSg	char(30)
SgStart	int
SgEnd	int
SgChr	varchar(2)
SgStrand	nchar(1)
CleavageSite	varchar(50)
CleaveStart	int
CleaveEnd	int
PriID	varchar(50)
MatID	varchar(50)
PAM	char(3)
DoenchScore	float
AzimuthScore	float
AzimuthScorev2	float

OverlappingSgRNAs

Variable	Type
SelectedSgID	int
OverlappingSgID	int

ControlSgRNA

Variable	Type
SgRNAName	varchar(50)
SgRNA	char(20)
LiteratureSgRNA	varchar(200)

InPool

Variable	Type
SgID	int
SgName	varchar(50)
OligoSeq	char(85)