Plant materials and RNA extraction
Garlic samples were collected form Pizhou, China (PW). PW garlic was cultivated in the test farm (Xuzhou city, Jiangsu province). The garlic was grown over 100 m2.
Using vernier caliper to measure garlic leaf length, width, thickness when garlic matures, and ensuring the selected material phenotype keep similar. Materials, after 0 h, 3 h, 6 h, 12 h wound dispose, were all obtained. The materials selection of garlic root, clove, gailic inner bud, garlic sprout were collected,then we assembled corresponding plant samples(Fig. 1a), immediately frozen in liquid nitrogen, stored at 80 ℃, until use. Total RNAs were extracted using the Plant Total RNA Isolation Kit according to its manual. By adding DNase I to the mixture to remove DNA contamination. Purified RNA was detected by 1% agarose gel electrophoresis. RNA was quantitatively detected by Implen nanometer photometer using Nanodrop with an RNA integrity number > 7.0.
Library preparation and transcriptomic analysis
The samples of approprite quality total RNA is prepared. The main process of library construction is as follows: magnetic beads containing Oligo (dT) are enriched with eukaryotic mRNA; The mRNA is randomly fragmented by Fragmentation Buffer; Acting as templates, the mRNA combined random hexamers primers to synthesize the first cDNA chain, then added the buffer, dNTPs, RNase H and DNA polymerase I to compound the second cDNA chain, and used AMPure XP beads to purify cDNA; Using purified double-stranded cDNA to repair and adding A-tails to connected to the sequencing beads. Then using AMPure XP beads to select the fragment size; Finally, the cDNA library was constructured according to PCR enrichione. When the library construction was completed, Qubit 2.0 and Agilent 2100 were used to detect the library concentration and Insert Size. To ensure the quality of the library, we use Q-PCR method to accurately quantify the effective concentration. After qualifying inspect library, we use HiSeq2500 for High-throughput sequencing. The reading length of sequencing was PE125. Clean Data that filter Raw Data is high quality data. It needs to be sequentially assembled. Trinity software will assemble Clean Data. In genetic identification and expression analysis, reads of different species were assembled together  (Fig. 1b).
Functional annotation and analysis
BLAST of NCBI was used to compare unigene sequences with databases, such as Non-redundant (NR) protein, Swiss-prot, Gene Ontology (GO), Clusters of Orthologous Groups of proteins (COG), EuKaryotic Orthologous Groups(KOG) and Kyoto Encyclopedia of Genes and Genomes(KEGG). Using KOBAS 2.0 to achieve the results of Unigene in KEGG of KEGG Orthology. We use HMMER and PFAM to compare the sequence after achieving annotation information of Unigene.
Differentially expressed unigene (DEGs) analysis
Bowtie compared the sequencing Reads of each sample with the Unigene library and estimated the expression level by RSEM based on the comparison results. The abundance of corresponding Unigene is indicated by FPKM value. The DEGs were screened with a criterions: FDR༜0.01 and FC (Fold Change) ≥ 2. The abundance values of transcripts were normalized. Using the MultiExperiment Viewer (version 4.9.0) to draw heat maps based on the transformation values. The figure showed that different columns represent different samples and different rows represent different genes .
Homology analysis and CDS prediction
TransDecoder software is used to compare the length of open reading frame, logarithmic likelihood function value and amino acid sequence with protein structure domain sequence in Pfam database. Predicted full-length sequences of the key genes, involved in the alliin synthetic pathway, were used for alignments.
Analysis of alliin contents
We prepare the materials of PW garlic roots, bulbs, inner bulb, garlic sprout, and wound leaves. Grinding the sample of liquid nitrogen, adding 4% of sulfosalicylic acid and ddH2O, 25℃ 30 min. and using centrifugal (12000rap, 20 min) to achieve supernatant. Then detect alliin content. To ensure the accuracy of the data,we measured the content of alliin at least three replicates ± standard error.
We adopted this effective method of Benjamini-Hochberg to correct p-value that hypothesis tests getting in the process of differential expression analysis. Finally, we uesed the adjusted p-values that FDR (False Discovery Rate). FDR is a key indicator for screening differentially expressed genes. It is an important methods to reduce the false positives of a large number of genes expression. Statistical analysis software is SPSS version 22.0. The method of comparing the differences uses ONE WAY ANOVA analysis of variance.