The T. fargesii plant sample used for cp genome sequencing was identified and artificially reproduced in the Wenchuan Botanical Garden of the Aba Tibetan and Qiang Autonomous Prefecture in Sichuan Province, China. The voucher specimen (Accession No. CP00002) was identified and deposited at the Herbarium of Neijiang Normal University (Neijiang City, China; Shixi Chen, saihei@foxmail.com). After collecting leaf specimens, they were stored at room temperature and packaged in 0.2 g of silicon dioxide. We used a modified cetyltrimethylammonium bromide (CTAB) method to extract total genomic DNA from fresh leaves (Porebski et al., 1997). Short reads of the T. fargesii cp genome were sequenced using a genomic library with an insert size of 260 bp. The products were prepared and sequenced on an Illumina HiSeq 2500 platform with approximately 436× coverage.

The sequenced reads were filtered using the Trimmomatic program (version 0.39) (Bolger et al., 2014). The filtered reads were assembled into the cp genome using NOVOPlasty (version 4.3.1) (Dierckxsens et al., 2017), and annotation was conducted using GeSeq with 3rd Party Stand-Alone Annotators from Chloë (version 0.1.0) and tRNA annotation from tRNAscan-SE (version 2.0.7) (Tillich et al., 2017). The annotation was followed by a manual check against the information from NCBI. Later, the sequences were deposited in the NCBI database (GenBank accession No. ON868919.1). The T. fargesii cp genome was visualized and plotted using Chloroplot (web-based tool: https://irscope.shinyapps.io/Chloroplot/) (Zheng et al., 2020).

Eight relevant Thalictrum cp genome sequences were downloaded from the GenBank database. Nine sequences, including T. fargesii, were analyzed to determine the degree of variation and sequence conservation within the same genus ( Supplementary Table 1 ). We compared the IR regions of all species using the web-based tool IRplus (https://irscope.shinyapps.io/IRplus/) to visualize the IR-SC boundaries and their gene orientations (Díez Menéndez et al., 2023). Sequence alignment was performed to identify the variation in sites between the nine Thalictrum cp genome sequences on mVISTA using the alignment program LAGAN (Mayor et al., 2000; Frazer et al., 2004). The sites were checked manually using BioEdit version 7 (Hall, 1999). The sites were subsequently analyzed, and the results are presented in Supplementary Table 2 .

Each amino acid has a minimum of one codon and a maximum of six codons owing to the simplicity of the codons. MEGA11 software was used to calculate the relative synonymous codon usage (RSCU) in the cpDNA of T. fargesii (Tamura et al., 2021).

SSR locus analysis was performed on the assembled T. fargesii cp genome sequence using the MicroSarellite identification tool (MISA) (Beier et al., 2017). The parameters were set as follows: 1–10, 2–5, 3–4, 4–3, 5–3, and 6–3, and the minimum distance between the two SSRs was set to 0 bp. The software package REPuter was used to conduct a long-repeat analysis and a minimum repeat size of 30 bp was assigned along a Hamming distance of 3 bp (Kurtz et al., 2001). The tandem repeat finder is an online program that detects tandem repeats using the default parameter settings (Benson, 1999).

We predicted RNA editing sites in the cp genome of T. fargesii using a convolutional neural network (CNN) model-based tool called Deepred-mt (Edera et al., 2021). We extracted the PCGs of this cp genome and inputted them into the Deepred-mt tool for prediction, considering threshold probability values greater than 0.9 as reliable results.

The natural selection that drives molecular evolution was analyzed for the Thalictrum chloroplast genome; thus, we calculated the ratio of nonsynonymous (dn) to synonymous (ds) substitutions called ω (dn/ds). The ω value is an indicator of natural selection of PCGs. Values ω > 1, ω = 1, and ω < 1 indicate positive, neutral, and negative/purifying selection, respectively. The stop codons of all 73 PCGs from these species were deleted and subsequently aligned, and the dn, ds, and ω values were calculated using MEGA 11 (Tamura et al., 2021).

To ascertain the phylogenetic position of T. fargesii within Thalictrum, nine species were analyzed, and Aconitum delavayi (NC_038097) was chosen as the outgroup. Sequence (full-length) alignment was performed using the MAFFT v7 plugin integrated into PhyloSuite v1.1.15 (Katoh and Standley, 2013). The dataset was refined using the Gblocks program with stringent parameter settings and then subjected to maximum likelihood (ML) analyses. Substitutional saturation was assessed using DAMBE version 7.0.68 (Xia, 2018) and phylogenetic trees were constructed using ML algorithms. ML phylogenies were conducted using RAxML v8.2.12, with 1,000 bootstrap replicates, and the GTRGAMMA model (Stamatakis, 2014).

After removing the adapter and low-quality reads, we found 9.4 Gb data for the T. fargesii cp genome. The cp genome was 155,929 bp in length, displayed a circular structure, and was divided into typical quadripartite regions, similar to most land plants ( Figure 1 ). The cp genome includes two 26,479-bp inverted repeats (IRA and IRB), and they separated the genome into a large single-copy (LSC) region of 85,395 bp and a small single-copy (SSC) region of 17,576 bp ( Supplementary Table 1 ). The overall GC content of the cp genome was 38%.

The size of the cp genomes of the Thalictrum ranged from 154,889 bp (T. thalictroides) to 156,103 bp (T. tenue). The cp genome of T. fargesii (155,929 bp) was the fourth largest and most similar to those of other species within the genus ( Supplementary Table 1 ). Moreover, we found the most extended LSC region (85,395 bp) in the T. fargesii cp genome, and the shortest LSC region (84,733 bp) in T. coreanum. In terms of count, the total number of genes found ranged from 128 to 134, and for PCGs, it ranged from 83 to 89. Nonetheless, except for T. viscosum (38 tRNAs), most species, including T. fargesii, possess 37 tRNA genes.

Annotation of the cp genome revealed 133 genes, including 88 PCGs (including six duplicated genes), 37 tRNA genes (including seven duplicated genes), and 8 ribosomal RNA genes. The identified genes were categorized into four groups: photosynthesis-related genes (n = 45), self-replicating genes (n = 58), other genes (n = 6), and unknown genes (n = 6) ( Table 1 ). Four pseudogenes have been identified in the genome. The IR regions carried two copies of ndhB, rps7, rps12, rpl2, rpl23, ycf2, trnI-CAU, trnL-CAA, trnV-GAC, trnI-GAU, trnA-UGC, trnR-ACG, trnN-GUU, rrn16, rrn23, rrn4.5, and rrn5. Of the genes identified, 21 contained introns and 16 (trnK-UUU, trnG-UCC, petB, petD, rpl16, rpl2, ndhB, rps12, trnI-GAU, trnA-UGC, rrn23, trnA-UGC, trnI-GAU, rps12, ndhB, and rpl2) contained one intron. The other five genes (two copies of rps16, trnV-UAC, and atpF) contained two introns ( Supplementary Table 3 ). Moreover, 10 genes were located in the LSC region, 10 in the two IR regions, and 1 in the SSC region. The largest intron (2,532 bp) was observed in trnK-UUU, and the smallest intron (199 bp) was found in rrn23. tRNA genes are distributed throughout the genome and are encoded by 61 possible codons (excluding the stop codon).

As shown in Table 1 , the PCG genes have different functions, including one acetyl-CoA carboxylase, six ATP synthases, one ATP-dependent CLP protease, five cytochrome b6, one cytochrome c biogenesis protein, one cytochrome b/f, one envelope membrane protein, two hypothetical chloroplast RF2, one megakaryocyte-associated tyrosine kinase, 12 NADH dehydrogenases, one N-terminal nucleophile amino hydrolase superfamily, seven photosystem I, 14 photosystem II, 25 ribosomal proteins, one ribulose-1,5-bisphosphate carboxylase/oxygenase large subunit, and four RNA polymerases. Each RNA had two copies and was equally distributed within the IR region. Duplicated tRNAs were not present in the LSC region. Moreover, rpl32, the rps19 fragment, one copy of ycf1, and infA were pseudogenes.

Base compositions of the LSC, SSC, and IR regions and codons at various locations were examined in the cp genome. The percentages of GC content in LSC, SSC, and IRs were 40.57%, 32.53%, and 43.23%, respectively, indicating that the percentage in the IR region was greater than that in the other regions. Moreover, the frequencies of adenine (A), thymine (T), guanine (G), and cytosine (C) contents were 30.46%, 31.12%, 18.84%, and 19.58%, respectively. The single-copy regions, LSC (59.43%) and SSC (67.48), had higher AT content than the repeat regions (56.77%) ( Table 2 ). The reduced AT content in the IR regions may be related to the presence of rRNA, which contains fewer AT nucleotides, contributing to genome stability and sequence complexity. Similarly, the CSD (protein-coding region) was rich in AT (61.24%), whereas the GC content was 38.73%. At the first, second, and third codon positions, each had a length of 25,038 bp and contained a range of 29.10%–31.44% A, 23.44%–37.25% T, 16.69%–27.35% G, and 14.61%–20.67% C contents, respectively. These findings disclosed that the cp genome had a higher AT than the GC content and that the codons preferred bases A and T.

The ratio of the frequency of a codon observed to that expected, termed “relative synonymous codon usage” (RSCU), was determined, which indicates codon usage bias in coding sequences. The protein-coding sequences (CDSs) of the T. fargesii cp genome were used to calculate RSCU via MEGA 11, and 64 codons were found. Three of these were stop codons and 61 encoded 20 amino acids. A single codon encodes methionine (Met) and tryptophan (Try), whereas the other amino acids are encoded by two to six codons ( Figure 2 ). Among the 64 codons, 31 had RSCU > 1 in the CDS of the cp genome, indicating that they were high-frequency and optimal codons. A or U was preferred as the third base of the high-frequency codon, as evidenced by the fact that 16 of these codons ended in U, 13 in A, 1 in G, and 1 in C. The most preferred codon was GCU, which encodes alanine (Ala) and has an RSCU value of 2. UUA, which encodes leucine (Leu) and has an RSCU value of 1.86, was the second-most preferred codon.

SSRs from cp genomes can be used to explore evolutionary relationships and population genetics, owing to their high polymorphism rates and consistent repetition (Xia et al., 2022). Most SSRs are A or T units that contribute to the AT richness of the chloroplast genome (Chew et al., 2023). Analysis of SSRs using the MISA web-based tool (Beier et al., 2017) revealed that the T. fargesii cp genome comprises 68 SSR loci. The most abundant were mononucleotides (35, 51%), followed by dinucleotides (20, 30%), trinucleotides (2, 3%), tetranucleotides (9, 13%), and pentanucleotides (2, 3%) ( Supplementary Table 4 ) ( Figure 3A ). Hexanucleotide repeats were absent. Among these repeats, 43.33%, 23.33%, and 25% were present in the LSC, SSC, and IR region, respectively ( Figure 3B ). In the cp genome of T. fargesii, the single-base repeat unit A/T (51%) is the most prevalent, followed by AT/TA (29%), AAAT/ATTT (7%), AATAT/ATATT (3%), and AAT/ATT (2%). The forenamed repeat units accounted for 92% of the total SSRs, whereas G/C repeat units accounted for only 8% of the total SSRs ( Figure 3C ). Among the SSRs, 21 were located in the intergenic region, 45 were within the CDS region, and the remaining 2 were miscellaneous.

Long repetitive sequences have proven to be valuable markers for comparative genomics, phylogenetics, genomic recombination studies, and plant evolution (Ivanova et al., 2017). According to our analysis, T. fargesii consisted of major repetitive sequences between 30 and 40 bp in length, including 30–32 bp, 33–35 bp, 36–38 bp, 38–40 bp, and ≥40 bp, accounting for 15, 4, 2, 4, and 3 repeats, respectively ( Figure 4A ). In total, the cp genome contained 27 long repeats, namely, 17 palindromic repeats (P), 9 forward repeats (F), 1 reverse repeat (R), and no complementary repeats ( Supplementary Table 5 ) ( Figure 4B ). Among the total repeats, 66.66% were in the LSC region, 29.62% were in the IR region, and only 3.70% were in the SSC region ( Figure 4C ). Only six genes (pafI, psaB, trnS-GCU, rpl16, ccsA, and ycf2) possessed long repeat elements, and ycf2 contained the highest number of repeat sites (n = 8, 29.62%). The remaining repeats were located in intergenic regions.

Centromeres are essential for chromosomal segregation and contain megabase-scale arrays of tandem repeats. These tandem repeats share common properties among species across different phyla (Melters et al., 2013). In total, 21 tandem repeats were identified in the cp genome of T. fargesii, 12 of which were dispersed in the gene spacer regions (rpoB/trnC-GCA, petN/psbM, petN/psbM, petN/psbM, psbZ/trnG-GCC, psbZ/trnG-GCC, atpB/rbcL, trnW-CCA/trnP-UGG, trnP-UGG/psaJ, trnP-UGG/psaJ, rps12/trnV-GAC, and trnV-GAC/rps1), and the remaining nine repeats were spread in the CDS region. One tandem repeat was found in rpl16 and each ycf2 gene had four repeats ( Supplementary Table 6 ).

Using Deepred-Mt, 242 high-confidence C-to-U RNA editing sites were successfully identified across 60 PCGs ( Supplementary Figure 1 ; Supplementary Table 7 ). Among these, ycf2 had the highest number of RNA editing sites (24), followed by ycf1 and rpoC2 (17 and 13, respectively), making them the top three genes for RNA editing. Additionally, both ndhB and psaB contained 11 spots, and accD, atpA, matK, ndhF, psaA, and psbA carried seven editing sites. In descending order, fewer editing sites were present in the PCGs, including six (psbC and psbD), five (ndhA and rpoB), four (atpI, ccsA, ndhD, ndhG, and ndhK), three (atpE, cemA, pafII, petB, and psbB), and two (atpB, atpF, ndhC, ndhH, rbcL, rpl16, rpl2, rpl36, and rpoC1). One editable site was detected in genes atpH, clpP1, ndhI, pafI, petA, petD, petL, psaI, psdJ, psbK, psbL, psbT, psbZ, rpl20, rpl23, rpoA, rps14, rps16, rps18, and rps7.

Significant contributions to variations within cp genomes among different species were observed, owing to the contraction and expansion of IR regions, which resulted in gene duplication, deletion, and pseudogene generation (Song et al., 2022). The locations of LSC/IR and SSC/IR junctions are regarded as indices of cp genome evolution (Zhang et al., 2013a). Figure 5 shows that the size of the cp genome within the Thalictrum was 154,889–156,103 bp; the lengths of the LSC and SSC were 84,733–85,507 bp and 17,470–17,657 bp, respectively; and the lengths of the IR regions were 26,272–26,521 bp. The LSC/IRB border is located within rps19 and extends 53–119 bp into the IRB region. Notably, the rps19 gene spanned 119 bp from the LSC to IRB in the cp genome of T. coreanum, whereas it stretched by only 100 bp in the same direction in the T. fargesii plastome. In T. thalictroides, the LSC/IRB border is flanked by an intergenic region of 28 bp between rps19 and the border, and the IRB/SSC border is between the partial ycf1 and ndhF genes.

In the SSC region, the ndhF gene was larger in T. thalictroides than in the other species. Moreover, T. foliolosum lacks ndhF. The ycf1 fragment gene crosses the IRB/SSC border as a pseudogene in T. viscosum, T. cirrhosum, T. balcalense, and T. fargesii. In contrast, the SSC/IRA junction was positioned on the full-length ycf1 gene in all the species. However, the size of the overlapping ycf1 gene at the SSC/IRA junction varies among the Thalictrum species. Remarkably, this gene enveloped 1,039–1,152 bp of the IRA region in all species. In addition, the IRA/LSC border was pointed between the rpl2 and trnH genes and 42–77 bp from the trnH region toward the LSC region. Notably, exceptions were observed in T. fargesii and T. coreanum, where the rps19 fragment was present at the edge of the IRA/LSC. The rps19 fragment ended at the termination of the IRA region in T. coreanum, whereas it was 2 bp extended toward the LSC in T. fargesii. These results demonstrate that the IR and SC borders of the cp genomes within the Thalictrum are primarily similar in terms of organization, gene content, and gene order.

Highly variable chloroplast genome sequences can be used to ascertain phylogenetic relationships between species and genera (Liu et al., 2018; Wu et al., 2018). To assess the degree of sequence divergence, we computed nucleotide diversity values (pi) to locate hotspots; hence, the results are presented for the entire T. fargesii cp genome in a plot of pi values ranging from 0 to 0.033 ( Figure 6 ). In the plot, the IR regions were shown to have a lower variability than the SSC and LSC regions. Remarkably, seven highly variable regions (pi > 0.02) were detected: three in the LSC (psbM-trnD-GUC, trnF-GAA-ndhj, and atpB), two in the SSC (ycf1 and trnR-ACG-rrn4.5), and two in the IRA (ndhD-psaC and rpl16-rps3). These regions may also be employed as high-potential fragments to distinguish between different Thalictrum species and may help in the development of molecular markers.

In this study, we analyzed the rates of nonsynonymous (dn), synonymous (ds), and dn/ds (ω) ratios among 77 PCGs common to Thalictrum, and the results are shown in Supplementary Table 8 . Among the 77 PCGs, the results of the statistical neutrality test indicated that the genes were relatively stable during the evolutionary process, and only 12 genes were found to have ω values less than 1 ( Supplementary Figure 2 ). To manifest, seven genes (atpF, matK, petG, rpl22, rps2, ycf1, and pafI) were under neutral selection (ω = 1), and four genes (atpE, ccsA, ndhG, and rpl20) had ω values of precisely 0.5, indicating that the genes were under purifying selection. The remaining gene, ndhF, had a ω = 0.33, suggesting that it was also under purifying selection.

The appearance of a complete T. fargesii cp genome provided us with access to sequencing data that could be used to investigate the phylogeny and, therefore, contribute to our understanding of evolutionary relationships within the Thalictrum. In the present study, nine species were analyzed to ascertain their phylogenetic position, and Aconitum delavayi (NC_038097) was chosen as an outgroup. ML analysis revealed a single phylogenetic tree with higher bootstrap values, which resulted in constructing one cluster from the complete cp genome of Thalictrum species divided into two clades ( Figure 7 ). Hence, Thalictrum was strongly supported as a monophyletic group [bootstrap support (bs) = 100%]. This result indicates that T. fargesii is closely related to T. tenue and T. petaloideum.

Regarding variation in gene content, we observed that pafI and pafII were found only in the cp genomes of T. fargesii and T. cirrhosum ( Supplementary Table 9 ). Moreover, in the case of pseudogenes, ycf1, infA, and rpl32 were identified in the T. baicalense, T. coreanum, T. thalictroides, and T. fragesii cp genomes, whereas the rps19 fragment (as a pseudogene) existed only in T. coreanum and T. fargesii, indicating significant variations. The remaining species in the genus showed a relative absence of pseudogenes in their cp genomes.

Variant sites are the most useful for identifying genes with significant variation and can facilitate the development of molecular markers (Li and Cullis, 2023). The whole cp genomes of the nine Thalictrum species were compared with those of T. fargesii, which served as a reference to investigate variant sites. Our results showed that 3.60% of the variant sites were present, and the conserved sites accounted for 96.40%, suggesting that these sequences were highly conserved within one genus ( Supplementary Table 2 ; Figure 8 ). Of the variant sites, 5,609 (3.60%) included 1.51% extragenic variant sites and 2.08% intragenic variant sites, indicating that the noncoding regions were highly conserved ( Figure 8A ). Within the intragenic sites, there were 1,500 (48%), 951 (30%), 550 (18%), 111 (4%), and 11 (0%) variant loci in the genes related to photosynthesis, self-replication, other, t-RNA, and unknown, respectively ( Figure 8B ). Furthermore, we calculated the frequency of variant sites within each category. Genes that contained more than 5% of the variant sites within each category included photosynthesis (ndhD, ndhA, pafI, petA, psbB, and ndhK; Figure 8C ), self-replication (rpl32, rpl33, rps18, rps19, rpl16, rps16, rpl20, and rpl2; Figure 8D ), t-RNAs (trnQ-UUG, trnC-GCA, trnK-UUU, and trnG-UCC; Figure 8E ), unknown (ycf1 and yfc15; Figure 8F ), and others (cemA, matK, accD, clpP1, and ycf1; Figure 8G ).

T. fargesii is of considerable importance in herbal medicine in Asian countries; however, taxonomic controversies need to be resolved, which may be possible from a detailed study of its plastome genome (Li et al., 2020). Such a systemic investigation and evolutionary analysis of T. fargesii has not been published previously. Therefore, our study reports the first whole cp genome composition of T. fargesii, which provides potential evidence via phylogenomics. Previous studies on other angiosperm groups have shown that chloroplast genomes are conserved or highly polymorphic (Abdullah et al., 2020, Abdullah et al., 2021). In the present study, the cp genome was highly conserved in terms of gene order, gene content, and intron number, following the findings of many other taxa, including Thalictrum (Tang et al., 2015; Yu et al., 2018; Xiang et al., 2022). However, in some taxa, such as Amorphophallus of Araceae, some genes were lost (Liu et al., 2019), and in others, such as Pelargonium (Chumley et al., 2006), the structure and gene order diverged from those reported here and in most other angiosperms.

Overall, the structure and organization of the nine observed cp genomes in Thalictrum, including T. fargesii, follow the typical structure commonly found in angiosperms, except for a difference in tRNA content in T. viscosum, which contains 38 tRNAs instead of 37 (Park et al., 2015; Daniell et al., 2016; He et al., 2019; Morales-Briones et al., 2019; He et al., 2021; Cai et al., 2022; Zhe et al., 2023). The total length, GC content, and gene composition were almost identical in the cp genomes of all studied species, and these highly conserved features might be due to a typical evolutionary passage in Thalictrum (Khan et al., 2020; Villanueva-Corrales et al., 2021; Feng et al., 2022). Compared to the LSC and SSC regions, the IR region was more conserved among the cp genomes.

In addition, the ratio of nucleotide substitutions (dn/ds) and their occurrence are frequently utilized as powerful tools to elucidate the evolution of protein-encoding genes and species adaptations (Kimura, 1989; Fay and Wu, 2003). It is noteworthy that the chloroplast genome of T. fargesii only contains neutral selection sites and purifying sites, suggesting that these genes have conserved functions throughout their evolutionary history (Khakhlova and Bock, 2006). These results offered further evidence that the cp genomes of the Thalictrum are conserved.

Next-generation sequencing was used to sequence the cp genomes of the ethnodrug species T. fargesii. The plastome of this species varies in terms of genomic structure, nucleotide diversity, structural alterations, and phylogeny with the plastomes of Thalictrum species (Park et al., 2015; He et al., 2019; Morales-Briones et al., 2019; He et al., 2021; Cai et al., 2022; Zhe et al., 2023). For instance, there is an apparent difference between Thalictrum plastomes in the IR-SC boundary areas, which causes variations in genome size (Xiang et al., 2022). IR-SC contractions or expansions into or out of neighboring single-copy regions are typically observed in angiosperm plastomes, and this process is a primary mechanism and an example of length variation in cp genomes (Kim and Lee, 2004; Yang et al., 2016; Zhang et al., 2016; Ye et al., 2018). Indeed, in our investigation, several genes, including ycf1, rps19, ndhF, ycf15, and ψrpl32, were worth examining more closely. For example, ycf1 is duplicated in the cp genomes of several members of the Thalictrum, with a larger copy (ycf1, 5,616–5,658 bp) positioned at the SSC-IRA boundary and a fragmented copy (ψycf1, 1,144–1,152 bp) located at the SSC-IRB boundaries, including T. fargesii (Xiang et al., 2022). Other taxonomic groups also contain the pseudogene ycf1 at the same junction (Henriquez et al., 2020; Mehmood et al., 2020). According to our findings, the shortest ycf1 gene was present in the cp genome of T. fargesii (5,520 bp), with more than 5% of the variant sites within the genus. Thus, ycf1 has enormous potential for use in developing molecular markers and has also been reported in Fritillaria and Pulsatilla (Bi et al., 2018; Li et al., 2018, Li et al., 2020).

Among such variations in cp genomes, molecular evolutionary phenomena including mutations, substitutions, selections, and random genetic drifts play a significant role. One of the molecular features known as “codon use bias” also helps unveil chloroplast and nuclear genome relationships and has been influenced by the GC content at codon locations in various organisms (Yang et al., 2018). Moreover, it is a critical feature for studying the functions of genes and contributes to its evolutionary relationship (Wang et al., 2023; Yang et al., 2023). We disclosed the presence of AT-rich intergenic regions in our present analysis, which is harmonized with the phenomenon that the overall GC content in the cp genomes of terrestrial plants is often low (Yang et al., 2017). Codon usage has a strong relationship with GC content in the codon position, and it is considered an important parameter in understanding evolutionary processes and selection pressure on genes (Zhang et al., 2013b). Our investigation revealed that of the 31 codons with RSCU values higher than one, 16 ended in U and 13 ended in A, indicating natural selection and mutation (Necşulea and Lobry, 2007). This finding suggests that T. fargesii prefers to employ synonymous codons with a third base, A or U, which is comparable to that of other Thalictrum species (Xiang et al., 2022; Zhe et al., 2023). This may be attributed to the abundance of A/T nucleotides in the angiosperm cp genomes (Abdullah et al., 2020). For instance, alanine and leucine were the most common codons in the T. fargesii chloroplast, which might be related to the abundance of A or T in the IR region (Chen et al., 2015), and is consistent with previous studies on the cp genomes of angiosperms (Orešič and Shalloway, 1998). Lauraceae exhibited a similar trend (Cao et al., 2023); however, the cp genome of Lespedeza showed comparable results (Somaratne et al., 2019).

Repeat sequences, particularly SSRs, are widely distributed in cp genomes, as microsatellite repeats play an essential role in gene expression, transcription control, and chromosomal construction, and display a high level of polymorphism (Qi et al., 2018; Huang et al., 2022; Xia et al., 2022). Our results support the hypothesis that cp-genome SSRs have substantial A/T bias, and this trend frequently occurs in many plants (Vieira et al., 2014; Wang et al., 2018). Similar to T. fargesii, mononucleotide repeats are the most abundant in Fritillaria, Lilium, and Epimedium (Rønsted et al., 2005; Zhang et al., 2016). Xiang et al. (2022) have reported 39 to 60 SSRs among 11 Thalictrum species; however, we have identified a total of 68 such repeats. In future genetic studies, we may leverage that SSRs could be used as a source of molecular markers and genomic barcodes for the T. fargesii cp genome. Previous studies have demonstrated that improperly combined repeating or sliding sequences can cause genomic rearrangements (Lu et al., 2017). Long repeat sequences promote cp genome rearrangement and increase the genetic diversity of a population (Guo et al., 2017; Cui et al., 2019). For example, the ycf2 gene provides a suitable resource for genomic studies in the cp genome of T. fargesii, similar to those in other plant species (Cauz-Santos et al., 2017; Shen et al., 2017; Liu et al., 2018). Owing to the presence of different repeats in genes, such as ycf1, ycf2, and rpl32, resulting in a pseudogenized gene or loss of functions, become a significant molecular feature of cp genome in plant evolution and identification (Santos and Pereira, 2018; Park et al., 2019; Miao et al., 2022).

In higher plants, RNA editing occurs as a codon-specific event for converting cysteine (C) to uridine (U) during posttranslational modification. Therefore, this molecular process is crucial for RNA maturation to prevent unwanted mutations and to act as a signal for adaptation (Chu and Wei, 2019; Somaratne et al., 2019). Numerous plants, including Arabidopsis, tobacco, pea, and tomato, have RNA editing sites (Wang et al., 2016). Thus, our findings on the significant C-to-U editing mechanism in the plastid genome of T. fargesii can aid in detecting missense mutations in its genes (Chen et al., 2018; Small et al., 2020).

Divergent regions in cp genomes can help to identify closely related species and reveal phylogenetic information (Dong et al., 2012; Du et al., 2017). These regions have been adopted to generate barcodes for plant identification in other families of plants, including woody plants (Tan et al., 2018; Yan et al., 2023). According to our findings, ndhD-psaC and rpl16-rps3 were found to be divergent hotspots in the cp genome of T. fargesii, and have also been reported in other studies on the plastome of Thalictrum (Xiang et al., 2022). The same phenomenon has been witnessed for numerous other taxa such as Lagerstroemia and Adrinandra (Cui et al., 2019; Dong et al., 2021; Nguyen et al., 2021; Ren et al., 2022; Yang et al., 2022). In addition, these mutational hotspots might aid researchers in studying population genetics and identifying Thalictrum species.

To evaluate the selection pressure on genes and genomes, dn and ds values are important markers in the cp genome. Selection pressure plays a vital role in the adaptation to different environments (Yang and Nielsen, 2000; Gao et al., 2019). Several genes in T. fargesii were found to have a lower dn/ds ratio, suggesting that they are likely to undergo disadvantageous non-synonymous substitutions and purifying selections (Sheikh-Assadi et al., 2022). In T. coreanum, T. foeniculaceum, T. foliolosum, and T. thalictroides, the ndhG gene was significant under positive selection, with a lower codon bias index, suggesting a lower expression level in vivo (Xiang et al., 2022). Moreover, they reported that atpF, petG, rpl20, and rpl22 were under positive selection. In contrast, in the present study, the ndhG and rpl20 were observed under purifying selection, and atpF, petG, and rpl22 were supported by neutral selection in T. fargessii. Thus, purifying selection might be a result of an evolutionary process that preserves the adaptive behavior of this species.

Interestingly, the pafI and rps19 genes in the plastome of T. fargesii exhibited more than 5% variation within the Thalictrum, which was consistent with the findings of Li et al. (2014). In addition, the pafI gene contains two mono-type (10 bp each) SSR repeats and a 39-bp forward repeat at one of its introns, similar to the findings elucidated by Kurt et al. (2023). With the meaningful characteristic architecture of the IRA-LSC, the distance between the rpl2 and trnH genes was markedly different from that of other species in the genus. Moreover, similar to the phenomenon observed in Fritillaria, the presence of the rps19 fragment at the IRA-LSC junction in the T. fargesii cp genome and its distance from trnH make this region a probable string for marker design (Li et al., 2014). Cumulatively, these findings provide molecular evidence for differentiating the ethnodrug plant T. fargesii from other species of Thalictrum.

The phylogeny and classification of Thalictrum have long been debated, and distinguishing T. fargesii plants from plants with similar morphological traits remains unclear (Xiang et al., 2022). Plastid genome sequences have proven useful for phylogenetic investigations of angiosperms (Jansen et al., 2007; Huang et al., 2014; Kim et al., 2015; Li et al., 2019). This approach is valuable for classifying organisms at lower taxonomic levels (He et al., 2012; Zhang et al., 2016). Previous phylogenetic studies of the genus Thalictrum have used traditional approaches to classify this genus that includes rpl16 intron and ndhC-trnV-UAC by Soza et al., (2012, 2013), and ndhA intron, trnLUAA-trnF-GAA, rpl32-trnL-UAG, and rbcL by Wang et al. (2019). Moreover, Xiang et al. (2022) found that 116 IGS regions had stronger support than 114 gene sequences and revealed that the closest phylogenetic relationship existed within T. minus and T. aquilegiifolium, and T. coreanum and T. thalictroides. Several studies have uncovered that close relationships were present between T. minus and T. tenue (He et al., 2019; Lin et al., 2021; Pu et al., 2022), T. tenue and T. petaloideum (He et al., 2019; Pu et al., 2022), T. minus and T. petaloideum (He et al., 2019; Pu et al., 2022), T. minus and T. aquilegiifolium (Cai et al., 2022; Michimoto et al., 2022; Xiang et al., 2022), T. viscosum and T. cirrhosum (Cai et al., 2022; Michimoto et al., 2022), and T. coreanum and T. thalictroides (He et al., 2021; Lin et al., 2021; Michimoto et al., 2022; Xiang et al., 2022) based on all sequences and whole cp genome. However, none of their circumscribed relationships have uncovered the T. fargesii status within this genus. The availability of complete cp genomes of T. fargesii, as investigated by the present authors, provides sufficient information to establish phylogenetic relationships among the same taxa. According to their results, T. fargesii is closely related to T. tenue and T. petaloideum, supporting the conclusion of an earlier study that T. baicalense is closely related to T. tenue, T. minus, and T. petaloideum (He et al., 2021). Interestingly, T. viscosum has been reported to be closely related to T. cirrhosum and T. foeniculaceum (Cai et al., 2022). Conferring with more Thalictrum taxon samples, our phylogenetic studies of cp genome sequences, reported for the first time, revealed significantly enriched phylogenetic resolution and provided robust inferences of intraspecific relationships. This finding expands the scope of future research on the correlation between these mutations in the cp genome in terms of genetic evolution and speciation in this species. Further studies should entail selecting more samples from wild populations based on their ecotypes and collecting more extensive data from both the nuclear and cp genomes to establish more detailed evidence of phylogenetics and species evolution.