To ensure that the flow rate reached at least 0.9 ml/min, used in the calculation for flow resistance and opening pressure (see Statistics and Figure 2), independent of shunt setting, while at the same time avoiding very high flow rates at low pressure settings, a maximum pressure was manually determined for each setting before starting the automatic protocol.
All analyses are based on the Drosophila melanogaster genome BDGP release 5. Sequencing reads in the raw fastq files were mapped using Bowtie1 [ 46] using default paired-end settings, a maximum insert size of 1000 bp for valid paired-end alignments (-X 1000) and reporting only uniquely mapped reads (-m 1).
Finally, to detect motifs enriched in all clusters, 3′ UTR sequences from 20 randomly chosen transcripts in each cluster (120 total) were compared against a scrambled background set in MEME (Bailey and Elkan, 1994), with the settings set at a maximum width of 8-nt and a search of the given strand only.
The settings used to run each method were: PSI-BLAST was run using NCBI NR database-added 534 query sequences, with the settings of a maximum number of 20 iterations (-j option), 0.0005 e-value threshold to include sequences for a profile construction at each iteration (-h option), and 1000 e-value threshold for returned alignment in the final iteration (-e option).
Molecular interaction networks explored by IPA tools, with the threshold settings of a maximum 35 nodes per network, revealed a total of 25 networks.
Tandem repeats were detected using the Tandem repeat finder (Benson 1999) with the recommended settings, and a maximum period size of 500 bp.
The conformational models of the different query molecules were calculated using omega (default settings, with a maximum of 400 conformations per molecule).
The reads were aligned with GSNAP [ 60] using default settings with a maximum of 4 mis-matches allowed against the Glycine max genome assembly and annotation v1.01 from Phytozome (JGI Glyma1.0 gene calls).
In order to generate networks in the core analysis, we used the settings of a maximum of 140 genes per network and 25 networks per analysis, because the networks for up to 140 genes allow for the possibility that the same network can include all focus genes [ 27].
For every subsequent analysis, the following settings were used: a maximum of 4.0 mismatches, 100 dinucleotide shuffles and a P-value threshold of 0.05.
For the database search the following settings were used: a maximum of two miss cleavages, oxidation as a variable modification of methionine, carbamidomethylation as a fixed modification of cysteine and trypsin was set as enzyme.
