Some 15 years ago, Gane (now Dr. Gane Ka-Shu Wong, Professor and iCORE Chair in Biosystems Informatics, University of Alberta, Canada) and I were staring at a set of plots and scratching our heads, wondering why there was a negative GC-content gradient when we aligned human transcripts from 5′ to 3′ to the genome. Until we had published several papers on other more interesting issues based on analyses of human genome sequences and variations [1], [2], [3] and found the same phenomenon from the rice genes a few years later [4], [5], had we realized the importance of this nearly universal feature albeit variable from bacteria to human [6]. Thinking along the line, we also did another exploratory experiment at University of Washington, taking the advantage of the Environmental Genome Project supported by NIH’s National Institute of Environmental Health Sciences, to re-sequence a couple hundred genes, of course including some interesting introns, especially those that are small in size (a median of 78 bp; also called minimal intron). The effort led to a realization of natural selection on functional sequence elements [7] in addition to just protein-coding sequences that can be evaluated with different methods [8], [9], [10].