Approximately half of the 18 million clusters contained at least one high-value compound that had high potency (with an IC50 or EC50 value of <10 μM), MeSH annotation, or Pharmacological Action annotation (Figure 6).
The remaining 35 siRNA targets were mapped to DNA repair pathways by GO annotation (Figure 1B, Table S2).
Additionally, these sequences were not found in the SwissProt database or had any other annotation (Figure 5).
In some cases the problem with the annotation was obvious as there was clear supporting evidence that the first intron should be deleted from the annotation (Figure 4).
The protein samples were separated by molecular weight using SDS PAGE and four gel slices from each lane were processed for LC-MS/MS analysis and protein annotation (Figure 4).
Considering that, in general, the longer the sequence the higher the chance of annotation (Figure S4) and number of GO terms recovered, we expect that a large number of these unannotated contigs (i.e., those longer than 500 bp) may correspond to novel or undescribed genes.
The majority of genes within each cluster were unique to that brain region cluster (last column of each cluster annotation; Figure S4, Figure S5, Figure S6, and Figure S7).
Different applications generate different type of files including sequence files in FASTA or FASTQ format (e.g. ORF or RNA prediction), TAB delimited text files (e.g. COG annotation, Figure 2c), graphic files, raw output files and so on.
Of 19,875 genes, 14,164 (71%) were represented by at least one transcript contig containing an almost complete CDS (> 90% in respect to that of the database entry used for annotation; Figure 2).
The nine modules are associated with hyphal morphogenesis, ion and small molecule transport, protein secretion, shifts in carbon utilization, stress responses, protein metabolism and catabolism, signal transduction, transcription-related processes, and other processes (proteins not belonging to the above-mentioned functional modules or lacking GO annotation, Figure 2).
