It can be challenging to use gene transcript level data by itself to empirically identify biologically relevant mRNA expression features in an unsupervised manner. In practice, genes with the greatest standard deviation (SD) or median absolute deviation (MAD) are routinely reported as interesting findings. However, these descriptive statistics often fail to identify genes that have been biologically activated or silenced. The most informative spacing statistic (MIST) identifies these interesting patterns in transcript level data by finding a pair of consecutive ordered observations that have a large difference (spacing) between them and are relatively near the center of the data. In a leukemia RNA-seq data set, 3 of the 4 genes with greatest MIST are involved in frequently recurring genomic alterations (TLX1, TLX3, TAL2). Also, for each of these 3 genes, the MIST-dichotomization of transcript levels was 88-97% concordant with the presence of alterations in the gene. In contrast, these 3 genes had MAD=0 and were ranked 29, 254, and 2190 by SD. These results indicate that MIST can more effectively identify biologically relevant genes than SD and MAD in some applications.