TL;DR (skip to the link below for a simple overview of how CODIS works).
It's complicated and as a writer, I feel self-conscious in trying to explain such a thing here on WS, but I'll give it a go. I have edited and re-edited this for brevity - there's so much more, but it's already long.
CODIS isn't a lab. It's a database. The FBI has strict (legal and academic) rules about which samples are allowed to be put against the DB as well as
which markers they use for this system, as we've discussed here before. WIDENING the way CODIS works would result in many, many people being investigated just because they were once in a house or were involved in the production of something - like the components of a house.
It would be a wild goose case, at the very least and VERY expensive, and disruptive to the lives of many people. Why stop at CODIS? CODIS wouldn't identify many, many murderers or violent criminals - those who have yet to be caught would not be in CODIS.
It would be unconstitutional, I believe, to run every single "sample" at every crime scene and the entire definition of "crime scene" would gradually change.
As it stands, people who are known to have been at a crime scene (investigators, roommates, neighbors, etc) are asked to give samples and the profiles are run for a process of elimination. Investigators have no choice about giving their DNA, but everyone else voluntarily give a sample (to match to crime scene samples). This is already a lot of labwork and a lot of samples. Individuals who are asked to give samples but don't are usually looked at more carefully. And it is those individuals who COULD be submitted to CODIS - except that CODIS rules (on which there are many pages of explanation) do not permit that.
CODIS is not designed for such massive use - it only matches DNA to felons, people in "the system" of prisons. By this, I mean it's not technically capable of doing this type of work - NOR does CODIS provide complete matches most of the time - but tracking down a felon is usually pretty easy if is a complete CODIS hit (which is not the same as a full profile - ever - because CODIS doesn't look at full profiles for matches, only.a carefully selected set of SNPs). CODIS also has a very clear notion of which SNP's are reliable for personal identification - of anyone; it is not about identifying typical citizens anywhere in the world - but about matching to US felons already in the prison system.
In this case, there are two "murder rooms," but also an entire house that was roped off as a crime scene. And several cars. I'm guessing they found fragments of hundreds of samples of DNA. CODIS wants the DNA it looks at to come from a carefully defined place within the crime scene (like a knife sheath - not the kitchen sink faucet handles, unless investigators believe the murderer used the sink).
The sample on the sheath was presumed to belong to the sheath owner/user (which is common sense). But investigators also knew there was a man matching DM's description (so NOT excluded) who lived within range of the murders, was a criminology student, and drove a white Elantra. IMO, they already had the name Kohberger on their list. No match turned up in CODIS - but a close match named Kohberger turned up through IGG. That lead to the trash collection at the parents' house.
Doing this type of investigation for EVERY stranger sample at a crime scene would result in many people's lives upturned or ruined, a severe slow-down of every single crime investigation, and of course, then Defense attorneys would claim it was an overly narrow, tunnel-vision focus on DNA to the exclusion of other obvious clues. This was a "party house," but it is not common sense to believe that there was a party that night (there wasn't) and that these murders were part of a typical party murder scene. It is clear from video that KG and MM were at a bar downtown and then at the food truck and then took a ride home then played on their phones and dialed up KG's former boyfriend, co-parent of Murphy. There was no party. KG's boyfriend had to be cleared (no point in using DNA databases there, at all - as it is expected that his DNA would be in the house - CODIS is not meant for that purpose).
And if Kohberger had NOT left the sheath, then all of this running of random people's DNA and upturning their lives by putting it through various databases would have been for naught. Because as far as we know, he left no other DNA at the scene - because he had carefully used everything he knew to try and avoid that - and almost did. Best laid plans of mice and men, etc.
I am guessing that many people gave DNA voluntarily to LE, who called for people to do so IIRC. I am guessing the nearby neighbor who said he'd been to the house gave his DNA. They also asked owners of white Elantras to voluntarily come forward and be cleared - and many did. But not that guy in Pullman (whom a WSU campus policeman had already noticed as owning a white Elantra and of course his white Elantra was registered to legally park at WSU housing - easy and standard thing for campus police to do).
LE in two states and presumably at the Federal level were all waiting impatiently to see if Chief Frye's appeal to white Elantra owners who had been out and about that night...to come forward. Many did. At least one did not. And then that Elantra owner's DNA shows up on a crime element (the sheath). Not just on a doorknob. So he did not come forward about his white Elantra (even though it's clear that everyone in his program was talking about these murders) but now admits he was indeed out driving it around that night.
I see why they submitted it to CODIS (probably not with any optimism - because I believe they already had Kohberger on their radar and knew he did not have a felony record; his misdemeanor theft conviction as a teen would not put him in CODIS - but the FBI could still see that conviction if they wished to - not sure it got to that point at the time).
Since the entire CODIS system is complex, there are many FAQ pages for LE to use, Vermont has a good one:
vfl.vermont.gov
It explains what can be used in CODIS pretty clearly, IMO.
