No Evidence of an Embedded Message in Human Ultraconserved DNA: A Bounded Computational Search with Stated Detection Limits

If one wished to leave a message that would last for hundreds of millions of years, where should it be placed? And how could it be made so widespread that someone would be likely to find it? DNA answers both questions: it is copied from generation to generation, and a message written into it would be present in every cell of every descendant. Such a message would survive only in sequence that natural selection does not allow to change. Earlier searches examined viral genomes and the genetic code; the ultraconserved elements, exceptionally conserved across vertebrate genomes, had not been formally tested. Here, 24 pre-specified statistics were applied to 2494 blocks identical in eight amniote genomes and, in a prospectively specified second stage, to the 481 canonical ultraconserved elements. Every two-bit and two-versus-two binary code was read for text, mathematics, pictures, periodicity, repeats and compressibility, and compared with dinucleotide-preserving shuffles (500 sets, extended to 2000 for the canonical elements); every detector first recovered planted signals. No message-specific statistic passed the corrected threshold in either set, or in the 213 canonical elements that share no sequence with the pilot. The structure exceeded that of shuffled sequence but not that of neighboring non-ultraconserved DNA. A weak picture-like signal was confined to elements that overlap protein-coding exons; in exploratory analyses its three-base periodicity was consistent with codon-related base composition, and its raster maxima were attributable to short tandem repeats in single elements. Implanted-signal experiments confirmed that the search would have detected text of 17 characters, 25 to 27 digits of a constant, or a 143-bit picture. Within the message classes, codes and sizes tested, these exceptionally conserved sequences carry no detectable message.