Correction

COMMENTARY. For the article “Extracting functional information from microarrays: A challenge for functional genomics,” by Michael Q. Zhang, which appeared in number 20, October 1, 2002, of Proc. Natl. Acad. Sci. USA (99, 12509–12511; First Published September 23, 2002; 10.1073/pnas.212532499), Fig. 1 appeared incorrectly. The locants for B and C were reversed. In addition, the locant for the description of D was omitted from the legend. The corrected figure and its legend appear below.

Fig 1.

Relations among different concepts in the SP-analysis method. (A) Expression profile matrix (table). t = (t₁, t₂, . . .) is the experimental condition index; in this example it indicates a set of time points. (B) Expression profiles (patterns). g₁ and g₄ are not strongly correlated directly, but both are strongly correlated with the correlated set (g_x, g₂). g_x, g₂ are the transitive genes interpolating the two terminal genes along SP₁ (see C and D); similarly, g_y is the transitive gene interpolating g₁ and g₅ along SP₂. (C) GO biological process tree. The Ps are process annotations for genes at a particular node. A gene may belong to more than one node (“multiple-function,” such as g₂). (D) Expression profile space. g_x is on the short path SP₁ terminated by the known genes g₁, g₂, and g₄ and hence is assigned a function of P₁,₁,₁,₁ (level L0) according to the GO tree in C; g_y is on SP₂ terminated by g₁, g₅ and is assigned a function of P₁,₁,₁ (level L1). g₁ is shared by both SPs and may be involved in both processes, which means the processes represented by SP₁ and SP₂ actually crosstalk to each other. The linked gene network can be formed by the subgraph SP₁+SP₂.