Ginseng

Starch was measured via an enzyme hydrolysis method. Starch was hydrolyzed into dual sugars by amylase, hydrolyzed into monosaccharides by hydrochloric acid, and finally determined by reducing sugar, which is converted to starch (Rose et al., 1991 (link)).
The contents of pyruvate in the sample were determined according to the methods of Lin et al. (1995 ). Protein was removed from the samples by TCA precipitation, and in the resulting sample, pyruvate reacted with 2,4-nitrophenylhydrazine. The product turned red in the presence of an alkali solution, and the intensity of the color change was measured by a spectrophotometer. A standard curve for calibration was obtained using sodium pyruvate as a reagent with a gradient of concentrations of pyruvic acid. Absorbance values were obtained to generate a standard curve to calculate the pyruvate concentration.
For glutathione (GSH), roots were ground in liquid nitrogen and homogenized in 1 mL 5% (w/v) m-phosphoric acid containing 1 mM diethylene triamine pentaacetic acid (DTPA) and 6.7% (w/v) sulfosalicylic acid. Root extracts were centrifuged at 12,000 × g for 15 min at 4°C. GSH contents were determined according to the methods of Kortt and Liu (1973 (link)) and Ellman (1959 (link)) with some modifications.
The ascorbic acid (AsA) content was determined according to Egea et al. (2007 (link)) with slight modifications. Ginseng roots were ground in an ice bath with 10 mL 5% metaphosphoric acid stored at 4°C, and then the final mix was homogenized by vortex. The final solution was maintained on the ice bath, in darkness, for 30 min and then centrifuged at 20,000 × g for 25 min at 4°C. Ascorbate was spectrophotometrically detected by measuring absorbance at 254 nm with a UV detector. For quantification of the compound, a calibration curve in the range of 10–100 mg kg⁻¹ was prepared from standard ascorbic acid. Results were expressed as mg 100 g⁻¹ FW.
Root extracts were centrifuged at 12,000 × g for 15 min at 4°C. The extraction and determination of ginsenosides was performed following the method of Yu et al. (2002 (link)).

Four-year-old plants of Jilin ginseng cv. Damaya growing in a production field were sampled at the fruit ripening stage (late July 2010). The plants were divided into 14 parts (named as tissues in this article) as shown in Figure S1, immediately frozen in liquid nitrogen and stored at –80 °C. Similarly, root systems were sampled from 5-, 12-, 18- and 25-year-old plants of Jilin ginseng cv. Damaya planted in forest on the Changbai mountains, immediately frozen in liquid nitrogen and stored at –80 °C. The ages of the plants were determined according to the time that they were planted and further confirmed based on their morphology.
In some gene expression studies, biological replicates followed by qRT-PCR confirmation were deployed. However, our (in preparation) and other35 (link) studies showed that digital gene expression profiling by high-throughput RNA sequencing is highly reproducible, with a correlation coefficient ranging from 0.90–0.98 (P = 0.000), no matter whether the samples were collected from different individuals growing in a replicate, in different replicates or in different years of an experiment, or no matter whether the experiment was carried out in environment-controlled or naturally field condition. Therefore, it is apparent that it is unnecessary to have biological replicates for digital gene expression profiling by high-throughput RNA sequencing. This conclusion is also supported by another digital gene expression profiling method (SAGE) facilitated by RNA sequencing using a traditional Sanger sequencer in which biological replicates were not usually applied29 (link)30 (link)31 (link). Moreover, the validation of gene expression by qRT-PCR has also been a concern. First, RNA alternative splicing is a universal phenomenon, which is leading to multiple transcripts (isoforms) per gene that are different in sequence domains. Analysis showed that there were 1–136 isoforms per gene, with an average of approximately 2 isoforms per gene (Figure S2). Furthermore, our (not shown) and other36 (link) studies showed that alternative splicing is tissue-specific, and the numbers and types of the isoforms for genes varied dramatically among tissues and across developmental stages. Therefore, while the expression of each isoform of a gene can be accurately measured by RNA sequencing, it cannot be done by either qRT-PCR or microarray because it is difficult to design qRT-PCR primers or microarray oligo probes to quantify expression of an isoform of a gene expressed in different tissues. This implies that the expression profiling results of both qRT-PCR and microarray may vary significantly, depending on where the qRT-PCR primers or microarray probes are designed in the target gene. Therefore, it is unreasonable to validate the gene expression results from RNA sequencing by qRT-PCR. Furthermore, because qRT-PCR is very low in throughput for gene expression profiling, only a few to dozens of genes could be analyzed to validate the gene expression results of high-throughput RNA sequencing. This number is less than 0.1% of the unigenes (e.g., 248,993) of a sample generated with high-throughput RNA sequencing, whose sampling size statistically is too small to meaningfully validate the gene expression results. Finally and importantly, as indicated above, the digital gene expression profiling by RNA sequencing itself is highly reproducible. Therefore, the samples collected from one biological replicate were analyzed and no qRT-PCR validation was performed for the digital expression profiling of the resulted isoforms so that the resources were used to sequence and analyze the transcriptomes of more tissues and more developmental stages in this study.

The root rot disease index (DI) was estimated by dividing the number of diseased ginseng roots by the total number of plants investigated. The severity of root disease observed in the pot experiment was determined by the presence of surface lesions, which were quantified on a scale from 0 to 4, with 0 representing no lesions and 1, 2, 3, and 4 standing for lesions that are <10%, 10%−33%, 33%−67%, and larger than 67% of the total area of the root. The severity of the disease at one sampling site was recorded as the DI, which was calculated as follows:
Si is the severity rating, Xi is the number of roots with the corresponding severity rating, and N is the total number of roots in one sampling site (Jiao et al., 2019 (link)).

KXS comprised Ginseng Radix, Poria, Radix Polygalae, and Acorus Tatarinowii Rhizoma mixed in a ratio of 3:3:2:2 by weight. KXS was obtained via extraction as described previously (Liu et al., 2014 (link)). Briefly, KXS was refluxed and extracted three times with a 10-fold quantity of 60% ethanol, each for 1.5 h. The extracts were dried under a vacuum and stored at −20°C for later use.