Possibly he means that you could design ARMS primers that only work with the normal or mutant sequences . There will be a sequence mismatch at the mutation position so you have to find an annealing temperature so that normal primers only amplify normal sequence and mutant primers only amplify the mutant sequence.

Another option is the DCAPS protocol where you introduce a new sequence change such that in combination with the mutant (or normal) sequence you have created a restriction enzyme site so that if you cut the pcr product then cutting at the right position shows that the mutant sequence has amplified

these techniques can be found at the below addresses

tetra arms pcr primer design

you can use this tool but finally you should optimise the primer practically, because primer designs by bioinformatic tool for ARMS_PCR some time fail in real discrimination of the SNPs in the lab.

https://snp.biotech.edu.lk/arms.php

dcaps or acrs dCAPS (wustl.edu)

Mutations is a diverse bunch; for TP53 it would commonly be loss-of-function, but that can be achieved by point mutations of by biallelic loss. For point mutations, you can simply sequence the coding region, but for allelic loss more complex methods are needed (e.g. adjusted coverage in whole genome sequencing). It you are working with established cancer cell lines, there is a good chance the mutations have already been mapped, but if you are working with patient samples, RNAseq is your best bet (or if on a budget, amplify the coding region of candidate genes, and hope a mutation will appear in the Sanger sequencing).

You can also try "melt-curve based genotyping" with a qPCR machine. You need primers that amplify a very short product (typically 80-120 bp). And you'd need a machine with "high-resolution" melting capability, meaning it can increase temperature by 0.1*C or less for each step.

This only works IF you can reliably see a difference in melting between the different possible alleles. You can only check one small region at a time and have 0 ability to detect novel variants.

Honestly, Sanger is cheaper & faster and you'll be able to get ~1000 bp in a read. Set up the PCR reactions in 96-well plates, use a multichannel (or liquid-handling robot), skip the agarose gel, and pay the sequencing center to do the cleanup & add the primers. Still cheaper than qPCR.

Sanger can also "phase" SNPS if they are close enough together to amplify in a single reaction. And you can look for rare or novel variants just as easily as known mutations.

Advice from someone who's seen a lot of ideas that look good on paper: if what you are looking for is *rare*, then your rate of false positives needs to be as close to 0 as possible. And your rate of false negatives *also* needs to be as close to 0 as possible.

There needs to be more rationale than "we could do [technique X]" when you know that Sanger will give you the data you need.