GenZedong

I completely agree with genetics being useless.

Your point about 'most generalized' is unfortunately not accurate. Perhaps in the past, there were for example major issues with the reference human genome which overrepresented people from western eurasia, and only have a handful of ethnic groups from Africa (recall, any two random people chosen from Africa will be more genetically distinct than any two other humans; Papua New Guinea might be an exception and other parts of Micronesia; at least to the best of my knowledge). Since for example there were many people from Amerika also sequenced, it only added to the bias towards western eurasians.

The current reference human genome is very very much improved over the initial attempts. Not perfect of course, however more specific databases and other references which are specific to other regions/groups/etc (really whatever grouping of genomes may be of interest), one which is not really related to ethnicity is the neanderthal genome, the Human Pangenome Reference tries to take care of the issues with any kind of bias by having a collection of genomes. There's also the African Variation Reference Genome.

What does this mean? The resolution is increasing, note earlier that if only western eurasians and certain ethnic groups from Africa, like Mbuti, Yoruba, etc. if one was not of that population, say a Nuxalk Nation individual, enough would be different that it would be hard to say maybe if they were closer to western eurasian population A or B. Which says nothing of how they relate to the people living in close proximity, say Inuit to the northish and Anishinaabe to the east.

That isn't the case now. There are single point mutations, groupings of a few bits of DNA, haplotypes (seminal paper here), which one can look at and say with confidence, "This genome belongs to a person who is very likely from this geographic region." This doesn't mean people don't have variation, only, the genome is conserved, i.e. it doesn't change all that much especially in some areas of the genome which are like super important. Such as the genes related to DNA Replication, making tRNA; stuff where if it wasn't there the cell couldn't grow and replicate at a high enough rate for the organism to be viable.

You can imagine there can be a test, like a pH strip, which only checks for one of several characteristic haplotypes or combinations of haplotypes if additional resolution is needed, and the result would be quite accurate. Enough to discriminate at least. Enough to arrest and test more thoroughly, which is only getting cheaper by the day.

You need to accept that this stuff is getting better, and outside of medical use, mapping migratory paths, etc. there need not be any application in a way which is discriminatory. Denying the effectiveness is denying scientific progress, maybe not now, but at least in the future. If we are to be dialectical materialists then it must be considered that a compound with a given causal history may have the ability to inform those who study it of its history, and, that ability will change over time. The incoming nefarious uses can only be countered if they are accurately anticipated. Humans exist and have existed in a specific physical space at specific times, like layers of sediment or the wear on a rockface, it can be analyzed and information of interest can be determined with great accuracy.

Look at this database, this population, Ecuador Cayapa has a 45.8% chance of having the haplotype DPA101:03-DPB104:02. You might say, "The sample size is only 183" and you would be right. What you aren't including, is the hundreds of thousands sampled (this is only for high quality data which meets the needs of this database, there is much much more; typically the data needs to be de-identified and meet a laundry list of requirements to even be collected, let alone shared) which represent thousands of different populations spread out across the globe. And, though the haplotype I gave had 'only' ad 45.8% chance, and if you check it out other populations have it too, you'll notice on the page there are three haplotypes which have a very high frequency. That makes it very precise. Here is the associated global map, there aren't very many places where you would find an individual with the given haplotype.

This doesn't include all the haplotypes which this population likely or very likely does not have. With that it's rather trivial to determine where an individual is likely from, than to infer the migratory pathways and history.

Sorry for the long post, I just wanted to convey that I knew a bit about what I was talking about. Not too much, others who do this for a living or have advanced degrees would know much more.

For some good news, look at this database which has data on which haplotypes are associated with adverse drug reactions.

A note on variation. There is a genome, some areas vary a lot, some don't as much. There are other 'omes. Omics captures the entirety of the field of this kind of computational biology. Transcriptome is more or less a time-series analysis of the level of RNA transcript of interest at given points in time. Epigenome is the epigenetic stuff, so like, there are molecules/complexes, the usual example is a methylated site. A methyl group is attached at some point along the genome and it changes how a gene is exposed. That changes how often it gets transcribed, how often a gene is expressed. You can't necessarily look at the genome to determine which sites are methylated. I don't remember the inheritance, it was sorta complicated and it's been a while. mtDNA, or mitochondrial DNA can also vary, there are some diseases which lead to a non-viable fetus or a baby which will not survive past a certain age because their body doesn't like produce enough energy needed.