CRISPR-Cas, and where the computing is

What it is

A bacterial immune system,repurposed

CRISPR is not an invention. It is a defence bacteria evolved against viruses, borrowed because it turned out to be programmable.

Adaptive immunity, in bacteria

Fragments of past viral infections stored in the genome as a memory of what to attack next time.

The name is the structure

Clustered Regularly Interspaced Short Palindromic Repeats — the spacers between repeats are the memory.

Programmable by accident

Because targeting is specified by RNA rather than by protein, you can point it anywhere by changing a sequence.

The mechanism

A guide, a protein,and a cut

Three parts, and the reason the technique spread so quickly is that only one of them has to change per target.

The guide RNAAbout twenty bases that match your target. This is the only part you change between experiments.

The Cas proteinThe nuclease that makes the cut, carried to the site by the guide and inert without it.

The PAMA short motif that must sit next to the target or nothing happens. It constrains where you can cut at all.

The repairThe cell fixes the break — error-prone, which knocks a gene out, or templated, which writes in your edit.

Earlier gene-editing tools needed a new protein designed and built for every target. CRISPR needs a new twenty-base RNA sequence, which you can order. That single difference is most of why the field moved from specialist to routine in a few years.

The variants

Cas9 is notthe only one

The system comes in families, and the choice between them is usually about what the cut looks like and where it is allowed to happen.

Cas9

The workhorse. Blunt double-strand break, NGG PAM, best characterised and most widely supported.

Cas12a

A staggered cut and a T-rich PAM, which opens up regions Cas9 cannot reach.

Base and prime editors

A disabled nuclease fused to an enzyme, changing letters without a double-strand break at all.

Cas13

Targets RNA rather than DNA, so the effect is transient and the genome is left alone.

Where the computation is

Designing the guide,and checking the damage

The wet lab does the editing. The two ends — choosing where to cut and finding out what was cut — are bioinformatics.

off-target search

$ makeblastdb -in genome.fa -dbtype nucl -out genome

$ blastn -task blastn-short -query guides.fa -db genome -outfmt 6

DesignBeforefind sites, score off-targets

SequenceAfterdeep coverage over the locus

AlignmentBoththe same tools as any NGS project

Guide design is a search problem: find twenty-base sites next to a valid PAM, then score every near match elsewhere in the genome, because a guide that also binds three other places will cut them too. Validation runs the other way — sequence the edited cells deeply and count what fraction carry the intended change against what fraction carry something else.

Both are alignment problems at heart, which is why a CRISPR project ends up with the same tools as any other sequencing project: an aligner, SAMtools, and a genome browser to look at the loci that matter.

The status

BLAST isbeing built

The image and its definition are being verified. When they are done, it deploys by name like the rest.

Ask for itSay so on the community tracker, and it moves up the list.Or build itOpen a pull request on the hub, following a published one.

Read it, then run itPrepaid and by the hour, on a machine that is yours about a minute after you ask.

Start a machine All notes