I like the cut of 'yer jib. I did this.
You are off by a factor of 10, unless you already have everything you need except a sequencer, and you plan on doing crap tests no one wants. You are gonna drop $300K just on sequencing validation samples. Almost all NGS tests are LDTs and require extensive validations, unless you get that one FDA-cleared one from PGDx. Then your per test cost is so high because you probably have to give 2/3 of your revenue to PGDx.
You are looking at 1-3 years for development, needing to set up the lab but also have sales and marketing, and need a way to deliver services to clients at little to no cost to the patient. How do you compete with FMI or national labs that give their tests away or pay ordering clinicians for "research studies"? Running 1 sample costs almost the same as running 30, except running 1 sample, even if paid full price, results in a $5K loss. How do you deal with low volume? Many payors are still not reimbursing. It takes balls of steel to pull this off.