DeepTrunc-cell predicts cellular CRISPR/Cas indel frequencies of matched and mismatched target–sgRNA pairs, including mismatch- and bulge-containing substrates, supporting guide spacer lengths ranging from GN16 to GN21.
DeepTrunc-in-vitro, the in vitro counterpart to this model for rare variant detection and enrichment, is also available.
Inputs must be in comma-separated format without indexes or headers, in the order of 'Target','sgRNA'.
'Target','sgRNA'
Example 35 bp Target: TCGAGATCGTAGCAAGCTAGCTAGCGGTTGACCTA (Full Target Sequence) = TCGA (5' Upstream Flank, variable-length) + GATCGTAGCAAGCTAGCTAG (Protospacer, variable-length) + CGG (PAM Sequence, 3 bp) + TTGACCTA (3' Downstream Flank, 8 bp)
TCGAGATCGTAGCAAGCTAGCTAGCGGTTGACCTA
TCGA
GATCGTAGCAAGCTAGCTAG
CGG
TTGACCTA
The Target sequence must be 35 bp total, where the length of the 5' Upstream Flank and Protospacers are based on the sgRNA's spacer length.
Example sgRNA: GATCGTAGCTAGCTAGCTAG (variable-length, starting with G)
GATCGTAGCTAGCTAGCTAG
The spacer length of target sequences and sgRNA can range from GN16 to GN21 (variable depending on the CRISPR–Cas variant).
Type of Model selects which nuclease variant scores the pair: DeepTrunc-HF1-cell, DeepTrunc-SP-cell, DeepTrunc-FnCas9-cell, DeepTrunc-FnCas9-en1-cell or DeepTrunc-FnCas9-en31-cell. All of them take GN16–GN21 spacers except DeepTrunc-SP-cell, which takes GN16–GN19.
Use df[['Target', 'sgRNA']].to_csv('input.txt', sep=',', index=None, header=None) when using Pandas to produce the input.
df[['Target', 'sgRNA']].to_csv('input.txt', sep=',', index=None, header=None)
Example input:
TCGAGATCGTAGCAAGCTAGCTAGCGGTTGACCTA,GATCGTAGCTAGCTAGCTAG
Outputs are provided as a comma-separated CSV file with column headers ['Target', 'sgRNA', 'Predict_Ratio'] (use df = pd.read_csv('output.csv', sep=',', header=0) to read the output file).
['Target', 'sgRNA', 'Predict_Ratio']
df = pd.read_csv('output.csv', sep=',', header=0)
