ArchIE method detects archaic ancestry without archaic reference genomes. Yoruba individuals average 7.97% ±0.6% archaic ancestry at 80% precision. Archaic ancestry in Yoruba depleted in selectively constrained genome regions due to purifying selection. Yoruba archaic segments total 258 MB with 2.1 MB at ≥50% frequency. Yoruba archaic ancestry not from Neanderthals despite slight Neanderthal matching or from Pygmy populations. Several Yoruba loci show archaic haplotypes at >60% frequency including NF1 tumor suppressor and HSD17B2 hormone regulator. ArchIE validates prior S* hits in Yoruba like XRCC4, TJP1, DUT. Europeans average 1.99% Neanderthal ancestry detected reference-free matching prior studies. Neanderthal ancestry elevated in Europeans at BNC2 and OAS loci. Archaic ancestry reduced in conserved regions across both Yoruba and Europeans indicating fitness costs. Yoruba archaic source diverged before YRI-CEU split and introgressed later. High archaic split times boost ArchIE accuracy due to more private mutations. Some high-frequency Yoruba archaic genes overlap positive selection signals. Method outperforms S* in simulations especially low admixture fractions. Ghost archaic admixture common in African populations like Pygmy matching prior findings.