Morphometric identification of thrips species (Figs 3: 1-6)
The taxonomic investigations revealed seven thrips species, viz., S. dorsalis Hood, S. oligochaetus (Karny), T. palmi Karny, T. tabaci Lindeman, F. schultzei (Trybom), Chaetanaphothrips orchidii (Moulton) and Asprothrips bimaculatus Michel & Ryckewaert on cotton (Table S2). Diagnostic characters and keys of the reported species are as follows.
Suborder Terebrantia
Family Thripidae
F. schultzei (Fig 3:1)
Diagnosis: An ocellar setae pair III that is as long as the side of the ocellar triangle and arises closely together between the anterior margins of the hind ocelli.The distance between the hind ocelli and the length of pair IV of the postocular setae. One pair of minor setae are present medially between the posteromarginal submedian setae on the pronotum, which has five pairs of major setae that are slightly shorter than the anteroangulars. Absence of campaniform sensilla on the metanotum. On tergite VIII, the posteromarginal comb is not developed.
S. dorsalis (Fig. 3: 2)
The body is yellow, and tergites III–VII have a brown median marking.The ocellar and postocular regions are closely striated. Pair III of the ocellar setae appear between the posterior ocelli, some distance behind the tangent of their anterior margins. As long as ocellar setae pair III, there are two pairs of post-ocellar setae. Pronotum closely striate, S2 posteromarginal setae longer than S1 Dark antecostal ridge and abdominal tergites.
S. oligochaetus
Diagnosis: The body is pale.Ocellar setae III situated between the posterior ocelli. Abdominal tergites with a pale antecostal ridge. Fore wing with 10 setae on the upper vein and 2 setae on the lower vein.
T. palmi (Fig.3: 3)
Diagnosis: The postocular setae pair I are slightly longer than ocellar setae III, and the ocellar setae pair III are small and arise outside the ocellar triangle.The metanotum has irregular longitudinal lines that converge at the posterior margin and transverse lines that curve anteriorly. The median setae start well behind the anterior margin, and campaniform sensilla are visible. The first vein of the forewing has three distal setae. Tergite VIII of the abdomen has a full comb, and tergite II has four marginal setae laterally.
T. tabaci (Fig.3: 4)
Diagnosis: Antennal segment V is not noticeably paler than antennal segment IV; the abdominal pleurotergites have closely spaced rows of regular, fine microtrichia; the tergite lateral margins have microtrichia on sculpture lines; and tergite IX has one pair of campaniform sensilla but the anterior pair is absent.
C. orchidii (Fig.3: 5)
Diagnosis: The fore wing is pale with basal and median cross bands of brown color, and the body is yellow. Eight segmented antennas. Ocellar triangle with ocellar setae pair III. The medial setae are positioned behind the anterior margin of the metanotum, which is freely reticulated. The first vein in the forewing has three distal setae, while the second vein has three to four. Abdominal tergites with craspedum; VIII with stippled area extending anteromesad from spiracle. Abdominal sternites with lobed craspedum, except medially on VII; median setae on VII arise ahead of posterior margin.
A. bimaculatus (Fig.3: 6)
Diagnosis: The abdomen is white with two brown patches on tergite VI, and the body is bicolored, with the head and thorax being brown.Thefore wing is brown with a white basal area. Within the ocellar triangle, ocellar setae pair III. Medially reticulated metanotum with campaniform sensilla present. With two robust apical setae on the hind tibiae. Tergites VII–VIII have a posteromarginal comb with a median row of small denticulations. Tergites are sculptured laterally and bear spines resembling microtrichia.
Diagnostic key to the reported species
1. Antennae 8-segmented…...………………………….……………………….………….................2
- Antennae 7-segmented………………………………….…......…………………………...............6
2. Metathoracic endofurca greatly enlarged, lyre shaped and reaching into the mesothorax……………………………………….…………………..… A. bimaculatus
- Metathoracic endofurca not greatly enlarged, U- or Y-shaped, and not reaching into the mesothorax……………………………………...............………………………………………….3
3. Tergite VIII with area of specialized sculpture extending anteriorly from each spiracle to antecostal ridge………………………..………….………………….. C. orchidii
- Tergite VIII without a large area of sculpture around the spiracles ................................................ 4
4. Lateral side covered by abdominal tergites without microtrichia; forewing second vein, homogeneous setal row…………………………..………………………F. schultzei
- Several microtrichia entirely cover the lateral side of the abdominal tergites, and the second vein of forewing has an uneven setal row .………………………………………….……………………5
5. Tergites of the abdomen with pale antecostal ridge...................................S. oligochaetus
- Abdominal tergites with dark antecostal ridge ................................................. S. dorsalis
6. Metanotal sculpture has median reticulations and abdominal pleurotergites include many rows of fine ciliate microtrichia; campaniform sensilla are missing........................................T. tabaci
- Metanotal sculpture is longitudinally striate, and there are campaniform sensilla present. Abdominal pleurotergites lack fine ciliate microtrichia.………….……………………………….T. palmi
Molecular identification of thrips species based on the mtCOI region
Based on homology search performed using BLASTx tool against the non-redundant (nr) nucleotide database of NCBI, the 50 different sequences were categorized into various four major species. The molecular identification revealed the existence of a thrips complex of three species: T. plami, T. tabaci and S. oligochaetus in cotton; two species, T. tabaci and S. dorsalis in onion; and a single species, T. palmi in brinjal. Among 40 isolates collected from diverse cotton areas, 38 were T. palmi (OL638468-82, OL672953-56, OL672960-61, OL672964-70, OL672974, OL672975, OL672977-84), one isolate from Hyderabad (Andhra Pradesh) was of T. tabaci (OL672959) and one isolate from Faridkot (Punjab) was S. oligochaetus (OL672951). All four isolates collected on brinjal from Solan (Himachal Pradesh) and Raichur (Karnataka) were of T. palmi (OL672971-73, OL672976). On onion, out of six isolates collected, five were T. tabaci (OL672952, OL672957-58, OL672962-63), whereas one isolate from Faridkot (Punjab) was S. dorsalis(OL672950) (Table S3). The identity percentage ranged between 97 and 100 with the NCBI nr database for all the samples with the respective thrips species.
Phylogenetic analysis
For the phylogenetic analysis of sequences from the present study, the T92 (Tamura 3-parameter) model of nucleotide substitution was found to be best suited for NJ analysis based on Akaike information criterion (AIC), which was carried out at 1000 bootstrapping (Fig. 4). Forty two isolates/ populations of T. palmi formed two distinct groups with 100% bootstrap support with majority falling in cluster 1 whereas, SLN4, SLN3 and SLN5 isolates from Solan, Himachal Pradesh, segregated in a distinct cluster 2. The sizeable number of T. palmi samples of cluster 1 have a mean intraspecific genetic distance of 0.003% (range; 0.002% - 0.013%) whereas cluster 2 have a mean intraspecific genetic distance of 0.016% (range; 0.012% - 0.019%) and the interspecific distance between the two clusters was 0.082%. The T. palmi isolates collected from cotton across all locations of northern India along with an isolate RAB 5 of brinjal collected from Raichur and DRW5B and DRW6 of cotton from Dharwad belonging to southern India, grouped into a sub clade of cluster 1 suggesting no distinct clades corresponding to host species or geographical location, whereas the T. palmi isolates of cotton from central (Nagpur, Surat amd Khandwa) along with Raichur cotton isolates from southern India formed a distinct sub clade of cluster 1. Similarly, the T. tabaci isolates from onion and cotton also clustered into a single group. The variability within the genus Scirtothrips seemed to be less with two species S. dorsalis from onion and S. oligochatetus on cotton clustering into a single distinct group supported by 100% bootstrap support. However, more isolates of each species might have led to a different conclusion.
The phylogenetic analysis of COI sequences from the present study and previously reported reference sequences from other countries or regions (Fig. 5) by the NJ method produced two major clusters, with cluster I comprising of populations from North, South, and Central India, while populations of temperate region from Himachal Pradesh falling in a distinct cluster 2. The cluster 1 further sub-grouped into two clades comprising majority of North India populations from Hanumangarh, Muktsar, Bathinda, Faridkot, Abohar, Fatehabad, Hisar, Sirsa, Jind and Delhi and few south Indian populations (Raichur and Dharwad) i.e. RAB5 (OL672976) from Brinjal and DRW5B (OL672975), DRW6 (OL672974) from cotton. Most of the populations from the central zone (Nagpur, Surat and Khandwa) along with two populations from southern India formed the second clade of cluster 1. The populations falling in cluster II included the isolates SLN3 (OL672971), SLN4 (OL672973) and SLN4 (OL672974) collected on brinjal from Solan (Himachal Pradesh) showed higher similarity with T. palmi isolate from Pakistan (KP845652) and an already reported Indian isolate (FM956392). The T. tabaci from Surat (OL672957, OL672958), Varanasi (OL672963), Faridkot (OL672952), Pantnagar (OL672962) from onion and Hyderabad (OL672959) from cotton formed a single cluster with maximum similarity with T. tabaci sequences from Israel (FN546149) and Pakistan (KP845584). The sequences of genus Scirtothrips form an outgroup with Thrips genus, further divided into two subclusters S. oligochaetus and S. dorsalis. The two isolates of Faridkot, one from cotton (OL672951), showed higher similarity with the S. oligochaetus isolate from Pakistan (KP871302) and another from onion (OL672950) had similarity with S. dorsalis from India (EU100984). The Dyothrips pallescens (KX697603) sequence was separated as an outgroup.
Global T. palmi haplotype diversity
Neighbor-joining phylogenetic analysis of mtCOI sequences based on 265 sequences of T. palmi (223 reference sequences and 42 from present studies) with two outgroup sequences of T. tabaci was performed. The results indicated that globally the T. palmi population can be divided into three major clades, where clade I (yellow colour) have isolates from India, Thailand, Pakistan, and Dominican Republic; isolates from Singapore, India, Pakistan, Taiwan, Japan, Dominican Republic, USA, UK, China, Thailand, and Switzerland fall in clade II (blue colour), whereas all the isolates from Indonesia grouped into a distinct clade (clade III- green colour) (Fig. 6). The Indian populations of T. palmi from the present study were also grouped into these two major clades, with the Himalayan thrips population in clade II, while the remaining population from the north, central, and southern zone falls in clade I.
Haplotyping based on COI sequence polymorphism analysis generated using DnaSP6 grouped T. palmi globally into 53 haplotypes with a haplotype diversity of 0.862. The Indian T. palmi represents a stable population with a long evolutionary history (Hd=0.8 and Pi=0.016) whereas the Indonesian T. palmi isolates which branched separately (Fig 7), have low Hd (0.2) and Pi (0.002) values which represent a recent bottleneck or founder even by several mitochondrial lineages (Table 1). The pairwise genetic distance (p-distance) of Indian T. palmi population with other countries ranged from 0.022 to 0.178 with highest similarity with the neighboring country Pakistan (p-distance: 0.022 and FST: 0.083) followed by Thailand (p-distance: 0.059 and FST: 0.539) whereas its most distinct with population from Indonesia (p-distance: 0.178 and FST: 0.0.953) (Table 2). Additionally, using a Minimum Spanning Tree to visualize the relationships between T. palmi haplotypes showed that the most prevalent haplotype, H5; HQ377269 (n = 67), is only found in India (Fig. 7). The following widespread haplotype was H10, EF117830 (n = 59), which was found in Pakistan, India, and Thailand. All the T. palmi populations from Indonesia formed a separate haplotype, H41; KF144140. The haplotype H3; AF378689 (n=27) was most prevalent globally, occurring in Switzerland, USA, Taiwan, Japan, Dominican Republic, UK, India and Pakistan. Other than these ~ 39 singleton clusters have been observed, however due to lack of sufficient sequences in each haplotype their validation needs further investigations. The Indian T. palmi populations segregated into 34 haplotypes, out of which two major haplotype were reported in the present study. These haplotypes were H43; OL638468 (n=29) which comprises isolates from Faridkot, Bathinda, Muktsar, Jind, Sirsa, Hisar, Fatehabad, Delhi and Hanumangarh of north zone and Dharwad and Raichur of the south zone. The second haplotype, H49; OL672954 (n=4) comprises isolates from locations in the central zone (Khandwa, Surat and Nagpur) and Raichur from the southern zone. The remaining 9 isolates formed singleton clusters with one to six polymorphic sites (Fig 7).
The Minimum Spanning Tree haplotype network of isolates from the present study only revealed that the populations from the northern, central and southern zones of India clustered into 12 haplotypes (Supplementary Figure S2); however, analysis with sequences from the global database revealed that the Indian populations segregated into 11 haplotypes. The haplotype H7; OL672967 (n=2) comprising isolates from Fatehabad and Hanumangarh locations of North India, segregated from the haplotype H1 with single polymorphic site. The pairwise genetic distance between sampled locations ranged from 0.001-0.075 (Table 3). The pairwise genetic distance between Himachal and other locations were all above 0.074, while the populations from Delhi, Punjab, Haryana, and Rajasthan seemed to be closely related, with pairwise genetic distances ranging between 0.001 and 0.002. The central zone population was closely related to the southern population with a pairwise genetic distance of 0.006, whereas with the northern population it ranges between 0.007 and 0.009 except for the Himachal population, which seemed to be highly diverse with a distance of 0.074. Similarly, a genetic distance of 0.004 to 0.005 was observed between the southern and north zone populations. The population from Himachal Pradesh here also showed a higher divergence, with a genetic distance of 0.075. The paired Fst values of Himachal Pradesh populations also supported its significantly high divergence compared to other geographical groups with negligible genetic flow among these locations, as indicated by higher pairwise genetic distance (p>0.05) (Table 3). The haplotype diversity (Hd) and nucleotide diversity (Pi) analysis of the present populations from diverse locations also suggested that the Indian populations are stable with a long evolutionary history or secondary contact between differentiated lineages (Hd = 0.584, Pi = 0.013). The populations of the South zone also followed a similar pattern with a stable population with a long evolutionary history (Hd=0.700, Pi=0.006). However, the North zone’s T. palmi populations represented divergence between geographically subdivided populations (Hd= 0.475, Pi=0.014) and the central zone exhibited a population bottleneck followed by rapid population growth and accumulation of mutations (Hd= 0.500, Pi=0.002) (Table 4).
Thrips population dynamics
The population dynamics under unsprayed conditions over the past 15 years clearly indicated a change in thrips incidence over the years. In the arbitrarily divided phase before 2010, the population of thrips remained below 5.0 thrips/ 3 leaves, with a single mean peak of about 5.0 thrips/ 3 leaves recorded from the 28th (9-15 July) to 30th SMW (23-29 July). This was followed by a declining trend in population, and by the 35th SMW, the incidence of thrips had almost disappeared in cotton. In contrast, the mean thrips population remained higher from June to the 39th SMW from 2011 to 2021 when compared to the mean population from 2005 to 2010. In the second phase, multiple peaks were observed, with a maximum thrips incidence observed at the 28th SMW, which was about 6 times higher than the mean population observed in the same SMW of the first phase. Thus the period from 2011 to 2021 witnessed an early higher thrips incidence fluctuating between 15 to 32.5 thrips/3 leaves from 25th to 30th SMW. Both phases witnessed a declining trend in thrips populations after the 30th SMW; however, in phase two, the mean population in the 34th SMW (20-26 August) declined almost to a level matching the peak thrips population of phase one (28th to 30th SMW). Phase two also observed a prolonged thrips incidence by 4 weeks from 35th (27 August to 2 September) till 39 SMW (24-30 September) compared to the first phase (Fig. 8) .