Suppose you have a soup of genetic material. What is the best way to identify something novel, as in, a fairly long sequence not previously identified that has

Suppose you have a soup of genetic material. What is the best way to identify something novel, as in, a fairly long sequence not previously identified that has much greater prevalence than baseline expectations at its closest genetic distance from a library?