PT - JOURNAL ARTICLE AU - Long Qian AU - Edo Kussell TI - Natural selection driven by DNA binding proteins shapes genome-wide motif statistics AID - 10.1101/041145 DP - 2016 Jan 01 TA - bioRxiv PG - 041145 4099 - http://biorxiv.org/content/early/2016/02/23/041145.short 4100 - http://biorxiv.org/content/early/2016/02/23/041145.full AB - Ectopic DNA binding by transcription factors and other DNA binding proteins can be detrimental to cellular functions and ultimately to organismal fitness. The frequency of protein-DNA binding at non-functional sites depends on the global composition of a genome with respect to all possible short motifs, or k-mer words. To determine whether weak yet ubiquitous protein-DNA interactions could exert significant evolutionary pressures on genomes, we correlate in vitro measurements of binding strengths on all 8-mer words from a large collection of transcription factors, in several different species, against their relative genomic frequencies. Our analysis reveals a clear signal of purifying selection to reduce the large number of weak binding sites genome-wide. This evolutionary process, which we call global selection, has a detectable hallmark in that similar words experience similar evolutionary pressure, a consequence of the biophysics of protein-DNA binding. By analyzing a large collection of genomes, we show that global selection exists in all domains of life, and operates through tiny selective steps, maintaining genomic binding landscapes over long evolutionary timescales.