Supplemental figures (titles and legends) Figure S1. Length distribution of VS16 454-ESTs. cDNA libraries were sequenced using the Roche/454 GS FLX system. EST sequences were filtered and clipped (Material and Methods). The resultant 1.5 million ESTs were binned in 10 bp intervals for plotting. Figure S2. Switchgrass genotype AP13 (a, b) and VS16 (c, d) node cultures on media with antibiotics and fungicide. Figure S3. Homology of switchgrass AP13 unitrancripts compared to reference genome. Unitranscript sequences of AP13 were BLASTed against transcript data of foxtail millet (E-value < 10-5). The numbers of top hits per foxtail millet gene transcript were plotted. Figure S4. Assembly of subsets of ESTs and pooling strategy for PviUTs. Six subsets of unitranscript sequences were produced separately. Major parameters used in assembly are included in round brackets adjacent to the programs (o, minimum overlap; p, minimum percent identity). Libraries and ESTs in black boxes were generated in this project. Others in the colored boxes were downloaded from NCBI without quality data. The length range and distribution of PviUTs are shown in the histogram. Figure S5. BLASTN E-values of PviUTs vs annotated foxtail millet transcript sequences Percentage number of total PviUTs were plotted over the top hit e-values (a). PviUTs with e-values between 1.0E-10 and 1.00 were further plotted using increment of 10-1, showing the e-value dropping off after1.0E-4. Figure S6. Functional annotation of PviUTs using CDD. Using the Conserved Domain Database (CCD) with BLASTX, significant domain footprints were assigned to 66% of all PviUT sequences. Figure S7. Switchgrass seed germination and time points for sampling. Figure S8. Stages of switchgrass inflorescence development. (a) Inflorescence meristem initiation, including rachis meristem (a1), primary branch meristem (a2), secondary branch meristem (a3), and floral primordium formation (a4); (b) Floret development; (c) Rachis elongation; (d) Panicle emergence. Figure S9. Dissection of switchgrass vascular fiber from an internode (a) and a magnified view (b). Figure S10. Signal intensity distribution of Affymetrix probe sets. Data accumulated at the Noble Microarray Facility were pooled and normalized before average signal intensity of each probe set was calculated. RNAs used were isolated from multiple organs exposed to various growth conditions or treatments. Five different Affymetrix cDNA chips were included. Log2transformed values were plotted. Figure S11. Number of probe sets detecting transcripts from various switchgrass organs. Supplemental tables (tiles) Table S1. Microbial sequence contamination and estimation of 454-EST-derived consensus sequences. Table S2. 454-EST assembly of VS 16 with different programs. Table S3. Newbler assembly metrics of 454-ESTs of VS16 with different parameter settings. Table S4. EST assembly and sequence resources used for the switchgrass unique transcript curation. Table S5. Estimated transcript and transcriptome coverage of switchgrass unitranscripts. Table S6. Alamo-AP13 × Alamo (non-AP13) tissues included in the gene expression atlas. Table S7. Summary of transcript sequence sources of probe sets and present/absent in the switchgrass gene atlas. Table S8 Most stably-expressed switchgrass genes. Table S9. Organ-enhanced genes. Table S10. Putative phenylpropanoid-monolignol biosynthesis genes and normalized average transcript levels. Table S11. Candidate phenylpropanoid-monolignol biosynthesis pathway genes in switchgrass and expression in 24 organs or tissues.