Gene networks in Drosophila melanogaster: integrating experimental data to predict gene function

Table 1 Datasets

Source	Dataset	Pass filter?	Genes	Relationships
Genetic interactions
FlyBase	All reported GIs	N/A	2,878	6,941
Protein-protein interactions
BIND, DIP, IntAct	Direct assay	N/A	935	1,234
DroID, BioGRID	High-confidence Y2H	N/A	4,543	4,590
	Positive Y2H	N/A	6,183	19,584
Microarray
Hooper et al. [100]	All conditions	Yes	10,460	3,289,275
Chintapalli et al. [102]	All conditions	Yes	10,054	3,618,216
Parisi et al. [92]	All conditions	Yes	9,922	5,656,854
Edwards et al. [99]	Line1	Yes	8,403	8,072,394
	Line2	Yes	8,296	8,118,665
	All conditions	Yes	0	0
Altenhein et al. [98]	All conditions	Yes	8,341	1,030,457
	Gof	No	0	0
	Lof	No	0	0
Hild et al. [97]	All conditions	Yes	8,214	677,746
Qin et al. [94]	All conditions	Yes	6,734	4,187,496
Tomancak et al. [103]	All conditions	Yes	6,288	2,626,310
Magalhaes et al. [59]	All conditions	Yes	5,718	1,102,629
De Gregorio et al. [57]	All conditions	Yes	5,698	1,561,265
	Bacteria	Yes	4,920	237,361
	Fungus	No	0	0
	Spaetzle	No	0	0
	Relish	No	0	0
	Spaetzle &relish	No	0	0
Sandmann et al. [101]	All conditions	Yes	5,474	1,238,924
Arbeitman et al. [61]	All conditions	Yes	4,354	1,769,479
	Embryo	Yes	4,126	1,271,286
	Larva	No	0	0
	Pupal	No	0	0
	Adult male	No	0	0
	Adult female	No	0	0
Sorensen et al. [96]	Heat	Yes	4,219	690,181
	No heat	Yes	4,083	701,546
	All conditions	Yes	0	0
Beckstead et al. [95]	Third instar	Yes	4,015	1,000,994
Estrada et al. [93]	All conditions	Yes	2,978	657,929
Wertheim et al. [58]	All conditions	Yes	2,280	551,684
Beckstead et al. [95]	Ecr	No	0	0
Li et al. [91]	All conditions	No	0	0

List of all datasets used in this study. The unit of data which we call a dataset is contained in the 'dataset' column. The filtering criteria apply to the microarray data as described in the Materials and methods section. The number of unique genes and functional relationships that a dataset contributes to the integrated network are listed. A '0' indicates that the dataset was not used for integration. There are two examples, Edwards et al. [99] (all conditions) and Sorensen et al. [96] (all conditions), where the dataset passed the filter but was not used in the integration. This is because all components in these experiments passed the filter criteria, but to remove redundant data, the subcomponent datasets were taken in favor of the dataset defined over the full set of conditions.

ISSN: 1474-760X