The problem with this approach is that DNA fingerprinting targets those sequences that vary most between between individuals (https://en.wikipedia.org/wiki/DNA_profiling). There are different techniques, and depending on the technique used, various features of the DNA are detected (restriction sites, PCR probes, etc).

Therefore, while it would theoretically be possible to mess up a particular sample, knowing the genomic sequence of the loci analyzed in the person from whom the sample originates and the particular fingerprinting assay to be used, the guide RNAs would have to be designed for each particular case and subject. In addition, the sample would have to be in a form in which the DNA is directly accessible, not embedded in tissues.

To mess up a specific sample, it would be much easier to just mess up the sample by mixing DNA from multiple donors or spiking in synthetic DNA fragments appropriate to the assay used.

Dear Annemarie Honegger , thank you very much for your answer.

I see your point: because the DNA fingerprint is almost unique to the individual, also the modification experiment has to be very specific. From what I read so far, the short tandem repeat sections seem to be a good starting point for the project. What do you think?

I agree that the modification cannot be an in-vivo experiment. In fact, the idea was to isolate a DNA sample, manifold it for i) the original DNA profile and ii) enough material for the modification. The guide RNA would be tailor-made and designed for cutting or exchanging marker loci such as a specific STR sequence.

The idea to mess up the sample is interesting. You are right, it would give a similar result, i.e a negative Q-K comparison. Yet, what you are suggesting is what I would call “noising”. The aim of this project, however, is a process I would call “ghosting” of DNA material. This might not make sense from a pure rational standpoint, but since the project is deeply rooted in a speculative and aesthetic design approach, the “ghosting” procedure is key and would mould the starting point for creation of the data sculpture (see .pdf-document attached for details).

So you are referring specifically to the STR method (https://en.wikipedia.org/wiki/STR_analysis) "A Short Tandem Repeat (STR) analysis is one of the most useful methods in molecular biology which is used to compare specific loci on DNA from two or more samples. A short tandem repeat is a microsatellite, consisting of a unit of two to thirteen nucleotides repeated hundreds of times in a row on the DNA strand. STR analysis measures the exact number of repeating units. This method differs from restriction fragment length polymorphism analysis (RFLP) since STR analysis does not cut the DNA with restriction enzymes. Instead, probes are attached to desired regions on the DNA, and a polymerase chain reaction (PCR) is employed to discover the lengths of the short tandem repeats."

Basically, the STR method uses primers flanking the STR loci to selectively PCR amplify the STR loci and compare their length. You would therefore have to alter the sites recognized by the PCR primers (highly conserved locations) rather than the STR regions themselves.

The problem is that the CRISPR/CAS9 method is not 100% effective, in the best published cases up to 86% (Article Efficient CRISPR/Cas9 genome editing with low off-target eff...

As a result, you would still have around 2% of the pairs of primer sites producing the correct PCR product, which could be compensated by 6 additional PCR cycles.

Dear Annemarie,

thank you for your answer. It's an interesting idea to alter the recognition site of the STR areas to prevent PCR amplification. As a result the STR would not show up in the profile. Yet, I agree with you, that a virtually 100% efficient alteration or cut-out would be required to completely suppress the signal. So that might be tricky. Still, I like the idea a lot, since it provides a different leverage position.

Regarding the effectivity of CRISPR/CAS9, I have the information that exchanging sequences is rather challenging, but cut-out should quite effective (86% is quite good for a knock-out, isn't it?).

Let's say we have a specific STR region and find out that there are 100 repetitions of a specific sequence with 5 nuleotides. Is it possible to design a gRNA to cut-out 4 of the nucleotide sequences so that the repetition is 96? Wouldn’t the resulting DNA profile then be different?

Best

Jo

The problem with acting directly on the STR is that these are highly repetitive sequences. If you have, for example, have a sequence of fifteen consecutive GAAA repeats, your 20NT guide RNA may bind just about anywhere in that stretch. And, while in a cell you have very few options of religating the cut DNAs, since you have only two of each chromosome, and the DNAs are additionally kept together by the protein scaffold of the chromosome, in vitro you have fare more options of ligating not the fragments of the same strand, but of different strands.

Dear Annemarie,

thank you for you answer, I highly appreciate your thoughts and input for this discussion.

So, the gRNA binds anywhere in the stretch and we have no control which of the GAAA repeats are cut, it could be the 1st, the 8th or the 10th one, right? For this specific case: does it matter? The results of the successful modification would be a new (modified) STR stretch with a different number of GAAA repeats. If this stretch is then "measured" for gene profiling, would the Q-K comparison give a "no-match"? Because, that's the only thing we want to see (see image attached). Or am I missing something, here?

I see your point with the in-vitro limitation. Do you think the experiment is impossible with isolated DNA? The issue with off-target effects would not be relevant for the research question (although it would be more elegant to have a clean modification) … as long as the gene profiles are different, we're good. Alternatively, is it feasible to cultivate specific human cells from the target sample for the experiment?

In the supplemental information to this paper Article STR sequence analysis for characterizing normal, variant, an...

If the gRNA can bind anywhere in the stretch and you are working in vitro, you would presumably get a distribution of different STR lengths, which may even be longer than the original length, since a DNA cut near the end of the STR can ligate with one cut near the beginning of the STR.

Ah, that's work by John M. Butler… I already contacted him and asked him about his opinion on this project. He suggested larger STR markers such as FGA, D21S11, or D18S51. Unfortunately, I don't have access to the article, but hopefully I'll get a response for the full-text request.

Regarding STR modification in vitro: That's a very interesting comment, I was not aware that an even longer version might be possible. In case a distribution of different STR-lengths are created: That would add several new peaks in the electropherogram, right? The original STR marker, however, would be almost gone, I reckon. That is, in case the modification was relatively efficient. The Q-to-K comparison would then be negative and the proposed "ghosting" achieved?