nkran/guido-old

Guide designer with microhomology predictor for MMEJ

guido-old's Issues

cut_sites to Pandas DF?

Should cut_sites be moved to Pandas DF or maybe Dask?

Test BWA for off-targets

• allows for finding off-targets with more mismatches

Resolving MMEJ subpatterns

Find a way to optimise removing shorter subpatterns.

PyVCF problem with reading reporting frequency

Extend to other PAMs

Support gRNA design for other Cas9 orthologs:

20bp-NGG - Sp Cas9, SpCas9-HF1, eSpCas9 1.1
20bp-NNG - Cas9 S. canis
20bp-NNGT - Cas9 S. canis - high efficiency PAM, recommended
20bp-NAA - iSpyMacCas9
21bp-NNG(A/G)(A/G)T - Cas9 S. Aureus
20bp-NNG(A/G)(A/G)T - Cas9 S. Aureus with 20bp-guides
20bp-NG(G/T) - xCas9, high efficiency PAM, recommended
20bp-NGN or GA(A/T) - xCas9, low efficiency PAM
21bp-NNN(A/G)(A/G)T - KKH SaCas9
20bp-NNN(A/G)(A/G)T - KKH SaCas9 with 20bp-guides
20bp-NGA - Cas9 S. Pyogenes mutant VQR
20bp-NGCG - Cas9 S. Pyogenes mutant VRER
20bp-NNAGAA - Cas9 S. Thermophilus
20bp-NGGNG - Cas9 S. Thermophilus
20bp-NNNNG(A/C)TT - Cas9 N. Meningitidis
20bp-NNNNACA - Cas9 Campylobacter jejuni
TTT(A/C/G)-23bp - Cas12a (Cpf1) Acidaminoc. / Lachnosp. - recommended
TTTN-23bp - Cas12a (Cpf1) Acidaminoc. / Lachnosp - low efficiency

Parse GTF annotations

Target haplotype enrichment

Check for haplotypes at the target site:

• Are there any population-specific haplotypes
• Highly abundant non-reference haplotypes
• Does targeting a specific haplotype effect the MMEJ outcome - out of frame deletions

Optionally target only selected features

Add option to target only exonic/intronic/intergenic/ ... sequences for a given region/gene.

Get rid of BioPython dependency

Add conservation score

Implement on-target efficiency score

https://genomebiology.biomedcentral.com/articles/10.1186/s13059-016-1012-2

Implement queue for processing

For processing the whole chromosome - chunk it and put it in a queue.

Create a python/conda package

https://github.com/pyscaffold/pyscaffold

Start writing tests

• - Choose testing library
• - Prepare test data
• - Prioritize parts of Guido to be tested first

Documentation

Optimize seed length in Bowtie

• 15 bases with no mismatches, and the rest with up to 3 mismatches (?)

Inspect variation in the sequence

Rank the guide lower if there is variation within the seed region or PAM

Design primers for homology arms

Add option to build genome

Bowtie needs to generate index in order to run the alignment for off-targets.
Enable guido-build genome.fa to build bowtie and fai index.

• inputs
  • genome fasta file (handle also .gz)
  • genome annotation (optional)
    • handle .gtf, .gff3 and .gz
  • genome base name
  • output data folder
• outputs
  • fai index
  • bowtie index
  • pickle with genome info
    • name
    • description
    • annotation filename
    • genome filename

Read about .2bit compression and see if it's worth using it to read sequences from. Depends also how it plays with GTF files.

Might be worth building on top of VectorbaseAPI branch.