The 'Dicty_Life' database and associated analysis constitutes a first attempt to link life history and developmental complexity to a mutant phenotype on a gene-by-gene basis. The strategy we pursued was to take movies of starving cells, from the very early events leading to cAMP wave propagation, through cell movement, aggregation, slug formation, and culmination; and to provide a quantitative representation of wave kinetics using a wavelet function to summarize wave dynamics, an approach that preserves wave frequency and amplitude as a function of time. It is an unbiased summary of an entire movie that can serve as a fingerprint of a gigabyte video data set.
The databaes currently contains dynamic records and wavelet representations for 2606 strains. The current release of Dicty_Life represents over 2200 of these strains that are described in details in Sawai et al. (2007) . Approximately 400 of these have been described in the literature or otherwise studied. The remainder come from the unbiased REMI screen at Baylor. Mutant phenotypes were confirmed by repeating the video analysis, sometimes as many as 6 times. Interesting strains were also cloned, their REMI insertion site verified by sequencing, and the phenotype recapitulated by transformation.
To gather video data (over 4 TeraBytes) on life cycles on this scale required the use of a robot, a snapshot of which is shown here. Clones were cultured in axenic media, plated on a thin agar surface in 6-well Costar plates, and scanned every few minutes, usually 6 plates per run (30 clones). To study wave propagation, averaged frames from each well were imaged by dark-field microscopy, and sequential images were subtracted for contrast enhancement. Slug histories were gathered in a similar fashion using bright-field optics. Additional experimental details may be found in Sawai et al. (2007) .
In addition to the life cycle videos linked to dictyBase, each entry is annotated according to standard and easily recognizable phenotypes - growth, waves, streaming, mound formation, slug phenotype, and culmination. Dicty_Life is searchable by three criteria related to each of the 6 annotated phenotypes - wild-type, mutant, and none. Because Dicty_Life is a work in progress, some of these data are incomplete (Undetermined). By entering a search term, a window will pop up that displays the number of hits by strain number, the mutant phenotype, and movies of the life-cycle and a wavelet portrait of the events leading up to late streaming. The wave movies contain both the raw video images and the frame-subtracted images.
The work was supported by a grant from the NIH/NIGMS (R01 GM063677) to Edward C.Cox. Large-scale insertional mutagenesis in the Adam Kuspa Lab at Baylor was supported by NIH /NICHD (PO1 HD39691). We thank members of the Dictyostelium Functional Genomics Group at Baylor, especially Christopher Dinh and Richard Sucgang for shipment, initial genotyping and PCR confirmation of mutant strains, Don Peoples and Richard Allan for robot construction, Rahul Sharma for database programming.
S. Sawai, X. Guan, A. Kuspa & E.C. Cox (2007) Genome Biology 8, R144 (provisional PDF). Highthroughput analysis of spatio-temporal dynamics in Dictyostelium.
M. Zhang, M. Goswami, S. Sawai, E.C. Cox & D.Hereld (2007) Mol. Microbiol. 65, 508-520. Regulation of G-protein-coupled cAMP receptor activation by a hydrophobic residue in transmembrane helix 3.
P.A. Thomason, S. Sawai, J.B. Stock, E.C. Cox (2006) Dev. Biol. 292(2), 358-70. The histidine kinase homologue DhkK/Sombrero controls morphogenesis in Dictyostelium.
S. Sawai, P.A. Thomason & E.C. Cox (2005) Nature 433, 323-6. An autoregulatory circuit for long-range self-organization in Dictyostelium cell populations.