Skip to main content
U.S. flag

An official website of the United States government

This site is currently in beta, and your feedback is helping shape its ongoing development.

Data from: The proportion of narrow fibers during early development predicts the diameter, fineness, and other quality traits of mature cotton fiber (<i>Gossypium</i> spp.)

Published by Agricultural Research Service | Department of Agriculture | Metadata Last Checked: February 12, 2026 | Last Modified: 2026-01-16
Cotton growth and samplingIn 2020, 161 historical Gh cotton accessions, along with 2 Gb accessions as comparators, were grown in an unirrigated field at the Sandhills Research Station, Jackson Springs, NC. The currently reported results include data collected on mature fiber quality and yield within a larger study [see (Billings et al., submitted) for further details and for seed sources]. The experiment had randomized complete block design with three replicates of two-row plots, and standard cultivation practices were used. Replicate 1 did not contribute to the currently reported results because it was flooded by rain at the beginning of the experiment when young fiber shape was phenotyped. For replicates 2 and 3, flowers were tagged on the day of flowering (0 DPA). Three days later, on 3 DPA, the ovules within one boll from each of four rows/plants were harvested. The boll wall was quickly cut away and the ovules were stabilized (fixed) in the field using our standard procedure (see further details below). At the end of the season a 25-boll sample was hand-harvested for fiber quality analysis from each of the rows/plots used for young fiber phenotyping, followed by bulk-harvesting to determine cottonseed and lint yield [see additional details in (Billings et al., submitted)]. The quality of mature lint fibers was analyzed by HVI and AFIS at Cotton Incorporated (Cary, NC) to generate the average values for each accession from replicates 2 and 3 that were used in the current data analysis.In 2023, young fiber shape was analyzed for five Gh accessions growing in five locations from Texas eastward in the USA. Four locations were reported in the 2023 Regional Breeders Testing Network (RBTN) multi-environment cooperative breeding trials (https://rbtn.cottoninc.com/), and the fifth location was the same Jackson Springs NC field that we used in 2020. The seeds for all five trials were produced in the field and distributed from a central stock to each location. Fields were planted in randomized complete block design with three replicates. The five accessions tested as each location were three non-transgenic, commercial, check cultivars (‘GhDP393’, ‘GhFM958’, ‘GhUA222’) and two elite breeding lines selected for high yield and improved fiber quality (‘GhOA23104’, ‘GhGA2017024’). Following the standard procedure, 3 DPA ovules were collected from 6 bolls per accession (2 from each of 3 field replicates at each location).We grew ‘GhDP90’ under controlled conditions at NC State over many years as part of the initial characterization of the two fiber shapes. These plants were grown from seeds derived from a single greenhouse-grown mother plant in a highly controlled Phytotron greenhouse as previously described (Pierce et al., 2019). Following the standard procedure, 3 DPA ovules were collected in the greenhouse. Alternatively, 1 DPA ovules were collected and cultured in vitro for two more days until ovules with attached fiber were collected on 3 DPA as previously described (Pierce et al., 2019; Graham et al. 2021).Fiber shape phenotyping and replicationOvules with attached fiber were fixed on 3 DPA using a low-toxicity, formalin-free fixative (A5472; Sigma‐Aldrich, St. Louis, Missouri, USA). This fixative preserved young fiber shape similarly to two others used in prior work that are either toxic or obsolete (Graham et al., 2022). Fixed samples were stored at 4°C prior to dissection and imaging. Phenotyping occurred at 3 DPA because prior experiments showed that the variation in ‘GhDP90’ fiber shape began to develop in chalazal fibers on 1 DPA and stabilized by 2 DPA (Graham et al., 2021). We predicted that most of the long ‘lint’ fibers of diverse accessions would have passed through this critical stage of differential shape formation by 3 DPA, which has been verified by further observations reported here. The short ‘fuzz’ fibers that coat Gh seeds typically initiate later than 3 DPA (Zhang et al., 2007) and had little or no impact on classifying the apical shapes of 3 DPA fibers that were extending outward from the surface of an ovule fragment (Stiff and Haigler 2016).Immediately before imaging, all the ovules in one fixed sample were examined under a stereomicroscope to choose two well-developed ovules. For each ovule, a chalazal fragment with attached fibers was removed with a fine scalpel and mounted in one well of a multiwell slide prior to semi-automated collection of stacked digital images in a Keyence BZ‐X810 imaging system (Keyence Corporation, Itaska, Illinois, USA) (Graham et al., 2022). For large data sets, computer vision was used to classify the 3 DPA fibers as narrow or wide. Smaller data sets were sometimes analyzed manually, and in this case the shapes of fiber apices with 6 - 10 µm diameter were cross-checked visually (for tapered or hemisphere appearance) to ensure reliability of the shape classifications. A circle was fitted by hand into the fiber apex to determine its diameter (Pierce et al., 2019; Graham et al., 2021).The comparative data for the apical diameter and average proportion of narrow fibers at 3 DPA for ‘GhDP90’ were generated manually during several years. Data sets 1-5 represent n = 2061, 720, 180, 180, or 180 fibers, respectively, from 3-4 bolls of different greenhouse-grown plants. The data in set 6 reflect n = 239 in vitro fibers from 6 ovule cultures within 3 independent trials. The data in set 7 reflect n = 719 fibers from a subset of the images used to produce computer vision results from the 2020 Jackson Springs NC field.For the 2020 field trial (Jackson Springs NC) of 161 Gh and 2 Gb accessions, each average narrow fiber proportion derived from computer vision analysis of 24 images representing 6 ovules from each of 4 bolls, one from each of two rows within two field replications. The average results represent 431 - 1130 fibers per accession (median = 725 fibers), which is typical of other results (Graham et. al., 2022). A subset of these accessions and images were also analyzed manually.For the 2023 multi-location cooperative breeding field trials in which we analyzed 5 Gh accessions at 5 locations, each average of narrow fiber proportion derived from computer vision analysis at least 21 images/ovules from 6 bolls. Providing perspective for future phenotyping, some of these 3 DPA samples were longer than optimum for imaging and classifying fibers by their apical shape (Graham et al., 2022). The apex of overly-long fibers was more frequently out of the imaging field and/or obscured within the mass of long, folded, fibers. When necessary, we additionally sampled shorter fibers on slower-developing ovules that are commonly present within one boll (Davidonis and Hinojosa 1994). In the future, sampling at 2 DPA may be preferred at some locations because young fibers elongate faster at warmer temperatures (Xie et al., 1993).

Find Related Datasets

Click any tag below to search for similar datasets

data.gov

An official website of the GSA's Technology Transformation Services

Looking for U.S. government information and services?
Visit USA.gov