3.1. Patient Characteristics and Groups
This study included 29 GISTs patients who regularly took IM orally and 11 GISTs patients who did not take IM. The included sample situation was displayed in Table 1. The average age of GISTs patients with IM or without IM were respectively 57.30 ± 8.42 and 54.12 ± 6.78, which demonstrated no significant difference. Similarly, the height, weight and sex were also recorded.
Table 1
Sample grouping situation
Characteristic | GIST group took IM(n = 29) | GIST group who did not take IM(n = 11) |
Mean age range(years) | 57.30 ± 8.42 | 54.12 ± 6.78 |
Mean height range(cm) | 161.71 ± 8.33 | 164.35 ± 7.93 |
Mean weight range(Kg) | 59.29 ± 11.52 | 54.61 ± 9.47 |
Male | 15(51.72%) | 6(54.55%) |
Female | 14(48.28%) | 5(45.45%) |
Dosage | 400mg qd | / |
The GISTs patients were divided into 4 groups, according to with or without IM treatment. Specifically, the first three groups (A, B and C) were received the IM 400mg/qd treatment. In detail, group A represented the IM plasma concentration ༞2000 ng/mL (n = 9); group B represented the IM plasma concentration ranging from 2000 ng/mL to 1100 ng/mL (n = 10) and group C represented the IM plasma concentration ༜1100ng/mL (n = 10). The last group D (n = 11) was a representative of IM missed patients. As shown in Supplementary S1, the difference between the highest IM concentration to the lowest IM concentration was more than 12 folds. This consequence reflected the huge individual differences when GISTs patients receiving the treatment of IM.
3.2. Influencing factors of gender, age and surgical operation with mean plasma concentration of IM
According to our previous literature analysis[4], the gender, age and surgical operation could be some controversial factors to influence the mean plasma concentration of IM. Herein, in this study, we conducted those factors to validate again.
Firstly, the plasma concentration of IM was detected. Consequently, the Shapiro-Wilk test for normality [31], Mann-Whitney U test for evaluating treatment effects in randomized elements [32], and independent t test for verifying statistically significant difference [33] were carried out to analyze the factors, which might be related with the mean plasma concentration of IM. The specific statistical data analysis of gender, age and surgical operation with Cmin of IM were supplied in the Supplementary data S1 to S3. As displayed in the Fig. 2A to C, the Cmin of IM in the female patients was higher than that of male patients, which was accordance with the report from a team of United Kingdom [34] (n = 93) and Switzerland [35] (n = 2478). The potential elements should be drawing to the mean body weight and clearance rate. As long-term results from the report of United Kingdom [34], when plasma concentration of IM was normalized for body weight, the differences in mean concentrations were no longer apparent. Herein, this consequence suggested that the higher plasma concentrations of IM in women could be partially explained by the lower in body weight, when compared with men. Furthermore, the clearance rate of women was higher than man from the report of Switzerland [35]. Due to these two reasons, the plasma concentration of IM was higher in women group. The age was demonstrated no significantly difference in our study, which was different with the Gotta et al. [35]. The Gotta et al. defined “young” as < 30 years of age and “elders” as people up to 70 years of age. The young group exhibited low IM concentrations compared with those of the elders (p < 0.05). However, firstly, the median age in our study was 57.30 ± 8.42. Specifically, the age of included GIST patients was from 40 to 78 years old. There were none of samples below 30 years old. Then, the sample capacity of our study was insufficient, which was limit by the COVID-19 [36]. As listed in the Fig. 2B, the surgical operation played important role in the plasma concentration of IM. The difference between gastric surgery and non-gastric surgery group was statistically significant. Moreover, partial gastric excision and total gastric removal were statistically significant in the gastric surgery group. Specifically, the concentration of IM in the total gastric surgery was lower than that of in partial gastric resection patients. Similarly, Yoo et al. and Hompland et al. [37, 38] reported that Cmin was significantly lower in patients who had previously undergone major gastrectomy than in those with previous wedge gastric resection or without gastric surgery. Obviously, decreased absorption of IM might be caused in part by the lack of gastric acid secretion in GISTs patients who had undergone major gastric resection. The gastric acid secretion was extremely important for IM absorption, because of IM tablets dissolved rapidly at pH 5.5 or less [39].
3.3. Correlation among mean plasma concentration of IM min with body surface area and some essential serum index
Linear mixed model analyses could display similar trends, although sometimes without reaching statistical significance. The body surface area was demonstrated by several reports [37, 40–42], which should be related with plasma concentration of IM. As displayed in the Fig. 2D to I, the body surface area was exhibited negative correlation with plasma concentration of IM (P = 0.043, r2 = 0.129). In yet other words, the higher dosage of IM should be given to the patients with smaller body surface area. Combined with ours’ and others’ consequence about the influence of body surface area, the fixed dosage of IM treated on the GISTs patients might not be suitable.
The laboratory data of ALT, AST, creatinine (CRE), total bilirubin (TBil), and albumin was also taken into consideration in this study. As demonstrated in the Fig. 2E to I, there was no significant correlation between ALT (P = 0.747, r2 = 0.003), AST (P = 0.667, r2 = 0.006), CRE (P = 0.905, r2 = 0.019), Tbil (P = 0.326, r2 = 8.468e − 4) and albumin (P = 0.583, r2 = 0.017) with plasma concentration of IM. The consequence of ALT and AST was consistent with the report of Yoo et al [37], while the result of CRE and Tbil were inconsistent with [ 40–42]. The reason should be associated with at least two hands. On the one hand, the number of samples among these above-mentioned studies was ranged from 25 to 89 GISTs patients. As we all known, the more included eligible samples, the statistics of data was more reliable. However, among these studies, the included samples were not very satisfied. On the other hand, the different ethnic population should not be ignored. The association of albumin with IMmin was controversial. It could be suggested that more IM would be bound to albumin in patients with higher albumin, resulting in higher total IMmin from the demonstration of Yoo [37], when the albumin exhibited active correlation with IMmin. Sometimes, the albumin was without reaching statistical significance with IMmin [38, 43]. The actual relationship of albumin with plasma concentration of IMmin was still unclear. Herein, not only large scale and multi-center trial of plasma concentration of IM related with influencing factors needed to be applicable, but also the mechanism behind the association should be explored, which would be promoted the rational usage of IM treating on patients with GISTs.
3.4. Principal Component Analysis (PCA)
PCA was belonged to unsupervised analysis assay. The advantage of PCA was that it could reduce the number of highly correlated metabolic features to a smaller set of principal components. This superiority made the PCA scores plots provide a visual description of the pattern described by the model that can be used for the identification of batch effects [44]. The results of PCA were illustrated in the Fig. 3.
The IM treated groups (A, B and C) could be distinguished from none-treated group D, especially in the ESI + mode, no matter in 2D or 3D plot. In the mode of ESI-, IM treated groups (A and B) were also tightly centralization when compared with group D. These consequences encouraged us for further statistical analysis.
3.5. Orthogonal partial least squares discriminant analysis (OPLS-DA)
The problem of insensitivity of variables with small correlation could be overcame by OPLS-DA model, which was the deficiency of PCA. Herein, we constructed the OPLS-DA model for the next data management. Firstly, we evaluate reliable and predictive ability of OPLS-DA model, which was supplied in the Supplementary S2. The parameters, R2X, R2Y and Q2 were all meet the standard. When the R2X, R2Y and Q2 were closer to 1, the more suitable and reliable for this model [45]. Meanwhile, the Q2 > 0.5 is considered as an effective model, Q2 > 0.9 belongs to an excellent model [46].
As exhibited in the Fig. 4, no matter in the mode of ESI- or ESI+, the Q2 was always more than 0.8, which indicated the models were all successful and biochemical changes between groups were clear. The abscissa of OPLS-DA plot represents the scores of major components. Therefore, the difference between groups could be judged from the direction of the abscissa. There was a clear separation between the IM treated groups (A, B and C) and group D. Concurrently, the samples from both groups tended to cluster in a concentrated manner, with a high degree of aggregation.
3.6 Differential metabolites screening
To distinguish the most important metabolites between the groups, we used the above analysis methods to digit the most potential differential metabolites.
3.6.1 Volcano analysis
After the OPLS-DA analysis, the specific numbers of statistical metabolites were conducted via volcano analysis. The variable importance in projection (VIP) of the OPLS-DA model can be used to preliminarily screen out the metabolites that differ among different groups. As exhibited in Fig. 5, the bigger the plot is, the value of VIP is higher. The value of VIP represents the influence intensity of corresponding metabolite differences in the classification discrimination of each group of samples in the model. Generally, the value of VIP ≥ 1 metabolite is considered to be significant metabolite. The red plots represent the significant up-regulation metabolites, while the green plots suggest the significant down-regulation metabolites. The rest, grey plots, recommend the insignificant metabolites.
Also, according to the Fig. 5, the total numbers of metabolites were 1020 and 2238 in the mode of ESI- and ESI+, respectively. This result suggested that the ESI- was less sensitive than ESI + when collecting IM metabolites. Herein, the mode of ESI + was carried out as the next analytical method.
It can be seen from Fig. 5B, D and F that there were totally 168 and 195 down-regulation metabolites, when compared with IM none-treatment group, respectively. Meanwhile, three were 125, 113 and 84 significant up-regulation metabolites, when compared with IM none-treatment group.
3.7 Biological information analysis
3.7.1 Cluster heat map analysis
Heat maps are one of the most widely used biological informatic graphic expression. Heat map is the representation of the metabolomics dataset with hierarchical clustering analysis (HCA) [47]. The graphical depictions of the metabolites are altered significantly across the different groups in the heat map in different colors. The heat map takes differential metabolites as the vertical axis, while the samples of model group or control group as the horizontal axis. The content of each differential metabolite in each sample could be directly seen according to the color depth. The darker the color is, the higher the content is [48]. As illustrated in Fig. 6A to C, the metabolites in the model group (A, B and C) and control group exhibited different behaviors (colors), which suggested these differential metabolites with good discriminating ability.
3.7.2 Pearson correlation coefficient
Pearson correlation coefficient was used to evaluate whether the relationship between the two groups was linear or nonlinear, or in other words, the strength of correlations [49]. Notably, 0 < Pearson correlation coefficient ≤ + 1 represents a positive correlation, while − 1 ≤ Pearson correlation coefficient < 0 represents a negative correlation. The greater the absolute value, the stronger the correlation [50]. We screened the top 50 differential compounds with the highest VIP value for Pearson correlation analysis. The each of top 50 compounds was calculated via Pearson correlation coefficient. Then a matrix graph was obtained, as displayed in the Fig. 6D-F, which displayed that red color represents a positive correlation between the two groups, while green represents a negative correlation between the two groups. The consequence of correlation would supply the information for narrowing the huge number of metabolites in order to dig the most potential differential metabolites.
3.7.3 Fold change
After qualitative analysis of differential metabolites, combined with the grouping situation of specific samples, the difference fold changes (FC) in the quantitative information of metabolites in each group were compared. The FC value represented the difference between the two groups and was calculated as the average of each individual peak area: (Mean value of peak area obtained from IM patients)/(mean value of peak area obtained from none IM treated patients). If the FC value was greater than 1 or less than 0.5, the metabolite was remained as significant one [51]. In order to make relative trend of change more intuitively, we took log2 FC to plot the histogram. Similarly, the bar charts in red were related with upregulation, while in green correlated with downregulation. As illustrated in the Fig. 6G to I, we screened the top 20 metabolites according to the log2 FC value, which belonged to meet the consequence of heat map and Pearson correlation. The range of most potential significant metabolites was further shrunk.
3.7.4 Violin box analysis
Given VIP > 1.0, FC > 2 or < 0.5, and p < 0.05 [52], the top 20 metabolites in treating groups (A, B and C) comparison with control group were identified, which exhibited in Table 2. Among them, the most obvious metabolite was N-Desmethylimatinib, which was the necessary and predictable in both D vs A, B and C. Indeed, the N-Desmethylimatinib was considered as major circulating active metabolite of IM [4].
Table 2
The top 20 differential metabolites identified by VIP, FC and P value.
Number | Groups | Code | Metabolites | RT/min | Regulation | P-Value | VIP | Fold Change | Ion Mode |
1 | D VS A | P0145 | L-(+)-Glucose | 0.85 | ↑ | ༜0.001 | 1.81 | 5.8566 | ESI+ |
2 | D VS A | P0189 | N-Desmethylimatinib | 4.14 | ↑ | ༜0.001 | 1.8 | 23.7459 | ESI+ |
3 | D VS A | P0925 | N-Demethylanhalidine | 3.48 | ↓ | ༜0.001 | 1.69 | 0.2708 | ESI+ |
4 | D VS A | P0635 | Securinine | 1.94 | ↓ | ༜0.001 | 1.64 | 0.4316 | ESI+ |
5 | D VS A | P0918 | D-Glucose | 0.84 | ↑ | ༜0.001 | 1.64 | 3.4835 | ESI+ |
6 | D VS A | P0851 | Palmitoyl-L-carnitine | 7.39 | ↓ | ༜0.001 | 1.61 | 0.3846 | ESI+ |
7 | D VS A | P0881 | D-sphingosine | 8.88 | ↓ | ༜0.001 | 1.6 | 0.4694 | ESI+ |
8 | D VS A | P0077 | Carnitine C18:1 | 8.12 | ↓ | ༜0.001 | 1.58 | 0.3234 | ESI+ |
9 | D VS A | P0079 | Carnitine C18:2 | 7.82 | ↓ | ༜0.001 | 1.58 | 0.3118 | ESI+ |
10 | D VS A | P0898 | 5'-S-Methylthioadenosine | 2.77 | ↑ | ༜0.001 | 1.54 | 2.2327 | ESI+ |
11 | D VS A | P0082 | Carnitine C18:3 | 7.39 | ↓ | ༜0.001 | 1.53 | 0.4196 | ESI+ |
12 | D VS A | P0324 | Carbetapentane | 6.85 | ↓ | 0.001 | 1.52 | 0.2518 | ESI+ |
13 | D VS A | P0123 | Iohexol | 1.68 | ↑ | ༜0.001 | 1.48 | 2.0734 | ESI+ |
14 | D VS A | P0331 | Ethylenediaminetetraacetic acid | 1.3 | ↓ | ༜0.001 | 1.44 | 0.4469 | ESI+ |
15 | D VS A | P0916 | 4-Hydroxyphenethyl alcohol | 2.87 | ↓ | 0.003 | 1.43 | 0.3129 | ESI+ |
16 | D VS A | P0853 | Isoferulic acid | 3.68 | ↓ | ༜0.001 | 1.41 | 0.4416 | ESI+ |
17 | D VS A | P0911 | (2R)-2-Hydroxy-2-methylbutanenitrile | 1.96 | ↓ | 0.004 | 1.37 | 0.4740 | ESI+ |
18 | D VS A | P0889 | L-Carnitine hydrochloride | 0.88 | ↑ | 0.003 | 1.37 | 3.2142 | ESI+ |
19 | D VS A | P0448 | 1-Palmitoyl-sn-glycero-3-phosphocholine | 8.5 | ↓ | 0.007 | 1.35 | 0.2636 | ESI+ |
20 | D VS A | P0739 | Octanoylcarnitine | 4.66 | ↓ | 0.007 | 1.28 | 0.2640 | ESI+ |
21 | D VS B | P0189 | N-Desmethylimatinib | 4.14 | ↑ | ༜0.001 | 1.98 | 16.7169 | ESI+ |
22 | D VS B | P0918 | D-Glucose | 0.84 | ↑ | ༜0.001 | 1.92 | 2.2934 | ESI+ |
23 | D VS B | P0927 | 1-sphingosine phosphate | 7.58 | ↓ | ༜0.001 | 1.9 | 0.4783 | ESI+ |
24 | D VS B | P0889 | L-Carnitine hydrochloride | 0.88 | ↑ | ༜0.001 | 1.78 | 3.1429 | ESI+ |
25 | D VS B | P0925 | N-Demethylanhalidine | 3.48 | ↓ | ༜0.001 | 1.76 | 0.2223 | ESI+ |
26 | D VS B | P0069 | 5'-Deoxy-5'-(Methylthio) Adenosine | 2.77 | ↑ | ༜0.001 | 1.7 | 2.0168 | ESI+ |
27 | D VS B | P0077 | Carnitine C18:1 | 8.12 | ↓ | ༜0.001 | 1.68 | 0.3136 | ESI+ |
28 | D VS B | P0079 | Carnitine C18:2 | 7.82 | ↓ | ༜0.001 | 1.67 | 0.3279 | ESI+ |
29 | D VS B | P0881 | D-sphingosine | 8.88 | ↓ | ༜0.001 | 1.67 | 0.3495 | ESI+ |
30 | D VS B | P0044 | Piperidone | 3.48 | ↓ | ༜0.001 | 1.66 | 0.4937 | ESI+ |
31 | D VS B | P0145 | L-(+)-Gulose | 0.85 | ↑ | ༜0.001 | 1.65 | 4.7190 | ESI+ |
32 | D VS B | P0851 | Palmitoyl-L-carnitine | 7.39 | ↓ | ༜0.001 | 1.65 | 0.4402 | ESI+ |
33 | D VS B | P0324 | Carbetapentane | 6.85 | ↓ | ༜0.001 | 1.64 | 0.2106 | ESI+ |
34 | D VS B | P0137 | L-Glutamate | 1.09 | ↓ | ༜0.001 | 1.63 | 0.4954 | ESI+ |
35 | D VS B | P0898 | 5'-S-Methylthioadenosine | 2.77 | ↑ | ༜0.001 | 1.63 | 2.1452 | ESI+ |
36 | D VS B | P0082 | Carnitine C18:3 | 7.39 | ↓ | 0.001 | 1.61 | 0.4322 | ESI+ |
37 | D VS B | P0331 | Ethylenediaminetetraacetic acid | 1.3 | ↓ | ༜0.001 | 1.51 | 0.4266 | ESI+ |
38 | D VS B | P0916 | 4-Hydroxyphenethyl alcohol | 2.87 | ↓ | 0.003 | 1.49 | 0.3205 | ESI+ |
39 | D VS B | P0911 | (2R)-2-Hydroxy-2-methylbutanenitrile | 1.96 | ↓ | 0.003 | 1.48 | 0.4452 | ESI+ |
40 | D VS B | P0853 | Isoferulic acid | 3.68 | ↓ | 0.002 | 1.46 | 0.4783 | ESI+ |
41 | D VS C | P0925 | N-Demethylanhalidine | 3.48 | ↓ | ༜0.001 | 1.95 | 0.2010 | ESI+ |
42 | D VS C | P0881 | D-sphingosine | 8.88 | ↓ | ༜0.001 | 1.92 | 0.3961 | ESI+ |
43 | D VS C | P0189 | N-Desmethylimatinib | 4.14 | ↑ | ༜0.001 | 1.89 | 8.8342 | ESI+ |
44 | D VS C | P0044 | Piperidone | 3.48 | ↓ | ༜0.001 | 1.84 | 0.4945 | ESI+ |
45 | D VS C | P0889 | L-Carnitine hydrochloride | 0.88 | ↑ | ༜0.001 | 1.78 | 2.9563 | ESI+ |
46 | D VS C | P0324 | Carbetapentane | 6.85 | ↓ | ༜0.001 | 1.77 | 0.2182 | ESI+ |
47 | D VS C | P0851 | Palmitoyl-L-carnitine | 7.39 | ↓ | ༜0.001 | 1.76 | 0.3943 | ESI+ |
48 | D VS C | P0898 | 5'-S-Methylthioadenosine | 2.77 | ↑ | ༜0.001 | 1.73 | 2.0964 | ESI+ |
49 | D VS C | P0331 | Ethylenediaminetetraacetic acid | 1.30 | ↓ | ༜0.001 | 1.72 | 0.4097 | ESI+ |
50 | D VS C | P0077 | Carnitine C18:1 | 8.12 | ↓ | 0.001 | 1.72 | 0.3521 | ESI+ |
51 | D VS C | P0079 | Carnitine C18:2 | 7.82 | ↓ | 0.001 | 1.7 | 0.3704 | ESI+ |
52 | D VS C | P0853 | Isoferulic acid | 3.68 | ↓ | ༜0.001 | 1.65 | 0.4093 | ESI+ |
53 | D VS C | P0916 | 4-Hydroxyphenethyl alcohol | 2.87 | ↓ | 0.002 | 1.64 | 0.2864 | ESI+ |
54 | D VS C | P0082 | Carnitine C18:3 | 7.39 | ↓ | 0.001 | 1.62 | 0.4345 | ESI+ |
55 | D VS C | P0371 | AG-1478 | 1.32 | ↓ | 0.001 | 1.57 | 0.4208 | ESI+ |
56 | D VS C | P0883 | Phytosphingosine;4-D-Hydroxysphinganine | 6.20 | ↑ | 0.005 | 1.49 | 5.0983 | ESI+ |
57 | D VS C | P0424 | 3-Cyclohexyl-6-(dimethylamino)-1-methyl-s-triazine-2,4(1H,3H)-dione | 2.66 | ↓ | 0.003 | 1.47 | 0.3402 | ESI+ |
58 | D VS C | P0448 | 1-Palmitoyl-sn-glycero-3-phosphocholine | 8.50 | ↓ | 0.013 | 1.47 | 0.3382 | ESI+ |
59 | D VS C | P0849 | Glycodeoxycholate | 6.90 | ↑ | 0.009 | 1.42 | 3.8801 | ESI+ |
60 | D VS C | P0739 | Octanoylcarnitine | 4.66 | ↓ | 0.0118 | 1.38 | 0.3266 | ESI+ |
As summarized in Table 2, there were glucose (P0145, P0918, P0925), securinine (P0635), carnitine (Carnitine C18:1(P0077), Carnitine C18:2 (P0079), Carnitine C18:3 (P0082) and mixture of those three (P0851, P0889), D-sphingosine (P0881), 1-sphingosine phosphate (P0927), phytosphingosine (P0883) and phosphoethanolamine (P0448) could be speculated into corresponding metabolism pathway, which we would like analyze with the help of KEGG pathway enrichment method as follow.
In this section, the distribution and probability density of top 20 differential samples in the Table 2 were inspected by violin box [53]. The thin black line extending from violin box represents the 95% confidence interval, the black bar in the middle of violin box represents the median value, and the outer shape represents the distribution density of the samples. As depicted in the Fig. 7A to C, the differential metabolites in the group D to A, B and C, exhibited different distribution patterns, which indicated the capacity of distinguish of these metabolites in the IM none treating or treating groups.
3.7.5 KEGG pathway enrichment analysis
KEGG (http://www.kegg.jp) is an encyclopedia of not only genes and genomes but also a professional tool for metabolites and nonmetabolic pathways [54, 55]. Pathways were considered significantly enriched, if p < 0.05, impact number of metabolite hits in the pathway > 1 [56]. The greater number of metabolite hits in the pathway with the lower p value, the more matching the significant pathway. The most influenced metabolic pathway was set as a pathway influence cut off value > 0.1 to filter for less important pathways. As displayed in Fig. 7D-F, the signal pathway as described included sphingolipid metabolism, drug metabolism (CYP 450), butanoate, caffeine metabolism and so on. The sphingolipid metabolism was with greatest commonness among those three groups, which impact numbers were above 1.75 with p value below 0.05. What follows is the metabolism map of sphingolipid metabolism, as described in Fig. 8G. The key metabolites such as D-sphingosine (P0881), 1-sphingosine phosphate (P0927), phytosphingosine (P0883) and phosphoethanolamine (P0448) were belonged to the metabolism pathway of sphingolipid, which was also existed in the Table 2.
3.8 Genotype of SNP
Five SNP (OCT1 1795 G > A(rs 6935207), OCT1 201 C > G(rs58812592), OCT1 1386 C > A(rs 622342), OCT1 1022 C > T(rs2282143), OCT11222 A > G(rs628031)) of 29 patients were performed by first-generation gene sequencing, and the typing results were shown in Fig. 8. As displayed in the Fig. 8, the results of genetic test demonstrated that the genotypes at OCT1 201C > G (rs58812592) were all CC (wild-type). The other four candidate SNPs of the OCT1 were with mutations. he genotypic distribution of the four SNPs were summarized in the Table 3. Herein, only the remaining four SNPs of genotypes were analyzed by Hardy-Weinberg equilibrium. Consequently, none of the genotypes showed deviation from the Hardy-Weinberg equilibrium [57] (P > 0.05). Then the relationship between genotypic SNP and IM plasma concentration were conducted through Mann-Whitney U test, which exhibited in the Table 4. Two genotypes at OCT1 1386C > A (rs 622342) was statistically significant (P < 0.05) with IM plasma concentration. The consequence of 1022 C > T(rs2282143)and 1222 A > G༈rs628031༉were similar with the report from Francis [58], which included the 73 patients of IM treated chronic myeloid leukemia patients.
Table 3
Genotype distribution of SNPs locus via Hardy-Weinberg equilibrium analysis
SNPs | Gene type | Number | Genotype frequency (%) | Theoretical value | PH−W |
OCT1 1022C > T(rs2282143) | CC | 22 | 75.86 | 22 | ༞0.05 |
TC | 7 | 24.14 | 6 |
OCT1 1386C > A(rs 622342) | AC | 5 | 17.24 | 5 | ༞0.05 |
AA | 24 | 82.76 | 24 |
OCT1 1222 A > G(rs 628031) | AA | 1 | 3.45 | 1 | ༞0.05 |
AG | 10 | 34.48 | 10 |
GG | 18 | 62.07 | 18 |
OCT1 1795G > A(rs 6935207) | GG | 8 | 27.59 | 7 | ༞0.05 |
AG | 13 | 44.83 | 15 |
AA | 8 | 27.59 | 7 |
Table 4
Relationship between IM plasma concentration and 4 genotypic SNPs
SNPs | Gene type | Number | Plasma concentration(ng·mL− 1) | P |
OCT1 1022C > T(rs 2282143) | CC | 22 | 1511.68 ± 842.38 | / |
TC | 7 | 1745.69 ± 1092.90 |
OCT1 1386C > A(rs 622342) | AC | 5 | 1488.81 ± 1375.69 | 0.005* |
AA | 24 | 1584.70 ± 800.84 |
OCT1 1222 A > G(rs628031) | AA | 1 | 300.34 | / |
AG | 10 | 1824.62 ± 1009.31 |
GG | 18 | 1496.13 ± 798.24 |
OCT1 1795G > A(rs6935207) | GG | 8 | 1699.52 ± 890.48 | / |
AG | 13 | 1861.02 ± 978.17 |
AA | 8 | 960.92 ± 400.60 |
P-Values were calculated from Mann-Whitney U test. * Represents P < 0.05. |