HairNet2: deep learning to quantify cotton leaf hairiness, a complex genetic and environmental trait
Abstract Background Cotton accounts for 80% of the global natural fibre production. Its leaf hairiness affects insect resistance, fibre yield, and economic value. However, this phenotype is still qualitatively assessed by visually attributing a Genotype Hairiness Score (GHS) to a leaf/plant, or by u...
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2024-03-01
|
| Series: | Plant Methods |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s13007-024-01149-8 |
| _version_ | 1797247254359179264 |
|---|---|
| author | Moshiur Farazi Warren C. Conaty Lucy Egan Susan P. J. Thompson Iain W. Wilson Shiming Liu Warwick N. Stiller Lars Petersson Vivien Rolland |
| author_facet | Moshiur Farazi Warren C. Conaty Lucy Egan Susan P. J. Thompson Iain W. Wilson Shiming Liu Warwick N. Stiller Lars Petersson Vivien Rolland |
| author_sort | Moshiur Farazi |
| collection | DOAJ |
| description | Abstract Background Cotton accounts for 80% of the global natural fibre production. Its leaf hairiness affects insect resistance, fibre yield, and economic value. However, this phenotype is still qualitatively assessed by visually attributing a Genotype Hairiness Score (GHS) to a leaf/plant, or by using the HairNet deep-learning model which also outputs a GHS. Here, we introduce HairNet2, a quantitative deep-learning model which detects leaf hairs (trichomes) from images and outputs a segmentation mask and a Leaf Trichome Score (LTS). Results Trichomes of 1250 images were annotated (AnnCoT) and a combination of six Feature Extractor modules and five Segmentation modules were tested alongside a range of loss functions and data augmentation techniques. HairNet2 was further validated on the dataset used to build HairNet (CotLeaf-1), a similar dataset collected in two subsequent seasons (CotLeaf-2), and a dataset collected on two genetically diverse populations (CotLeaf-X). The main findings of this study are that (1) leaf number, environment and image position did not significantly affect results, (2) although GHS and LTS mostly correlated for individual GHS classes, results at the genotype level revealed a strong LTS heterogeneity within a given GHS class, (3) LTS correlated strongly with expert scoring of individual images. Conclusions HairNet2 is the first quantitative and scalable deep-learning model able to measure leaf hairiness. Results obtained with HairNet2 concur with the qualitative values used by breeders at both extremes of the scale (GHS 1-2, and 5-5+), but interestingly suggest a reordering of genotypes with intermediate values (GHS 3-4+). Finely ranking mild phenotypes is a difficult task for humans. In addition to providing assistance with this task, HairNet2 opens the door to selecting plants with specific leaf hairiness characteristics which may be associated with other beneficial traits to deliver better varieties. |
| first_indexed | 2024-04-24T19:55:46Z |
| format | Article |
| id | doaj.art-307372dac3e741bca13f82732c72ebb2 |
| institution | Directory Open Access Journal |
| issn | 1746-4811 |
| language | English |
| last_indexed | 2024-04-24T19:55:46Z |
| publishDate | 2024-03-01 |
| publisher | BMC |
| record_format | Article |
| series | Plant Methods |
| spelling | doaj.art-307372dac3e741bca13f82732c72ebb22024-03-24T12:21:47ZengBMCPlant Methods1746-48112024-03-0120111910.1186/s13007-024-01149-8HairNet2: deep learning to quantify cotton leaf hairiness, a complex genetic and environmental traitMoshiur Farazi0Warren C. Conaty1Lucy Egan2Susan P. J. Thompson3Iain W. Wilson4Shiming Liu5Warwick N. Stiller6Lars Petersson7Vivien Rolland8Data61, Commonwealth Scientific and Industrial Research OrganisationAustralian Cotton Research InstituteAustralian Cotton Research InstituteAustralian Cotton Research InstituteAgriculture and Food, Commonwealth Scientific and Industrial Research OrganisationAustralian Cotton Research InstituteAustralian Cotton Research InstituteData61, Commonwealth Scientific and Industrial Research OrganisationAgriculture and Food, Commonwealth Scientific and Industrial Research OrganisationAbstract Background Cotton accounts for 80% of the global natural fibre production. Its leaf hairiness affects insect resistance, fibre yield, and economic value. However, this phenotype is still qualitatively assessed by visually attributing a Genotype Hairiness Score (GHS) to a leaf/plant, or by using the HairNet deep-learning model which also outputs a GHS. Here, we introduce HairNet2, a quantitative deep-learning model which detects leaf hairs (trichomes) from images and outputs a segmentation mask and a Leaf Trichome Score (LTS). Results Trichomes of 1250 images were annotated (AnnCoT) and a combination of six Feature Extractor modules and five Segmentation modules were tested alongside a range of loss functions and data augmentation techniques. HairNet2 was further validated on the dataset used to build HairNet (CotLeaf-1), a similar dataset collected in two subsequent seasons (CotLeaf-2), and a dataset collected on two genetically diverse populations (CotLeaf-X). The main findings of this study are that (1) leaf number, environment and image position did not significantly affect results, (2) although GHS and LTS mostly correlated for individual GHS classes, results at the genotype level revealed a strong LTS heterogeneity within a given GHS class, (3) LTS correlated strongly with expert scoring of individual images. Conclusions HairNet2 is the first quantitative and scalable deep-learning model able to measure leaf hairiness. Results obtained with HairNet2 concur with the qualitative values used by breeders at both extremes of the scale (GHS 1-2, and 5-5+), but interestingly suggest a reordering of genotypes with intermediate values (GHS 3-4+). Finely ranking mild phenotypes is a difficult task for humans. In addition to providing assistance with this task, HairNet2 opens the door to selecting plants with specific leaf hairiness characteristics which may be associated with other beneficial traits to deliver better varieties.https://doi.org/10.1186/s13007-024-01149-8Deep learningNeural networkMachine learningPhenotypingTrichomeCotton |
| spellingShingle | Moshiur Farazi Warren C. Conaty Lucy Egan Susan P. J. Thompson Iain W. Wilson Shiming Liu Warwick N. Stiller Lars Petersson Vivien Rolland HairNet2: deep learning to quantify cotton leaf hairiness, a complex genetic and environmental trait Plant Methods Deep learning Neural network Machine learning Phenotyping Trichome Cotton |
| title | HairNet2: deep learning to quantify cotton leaf hairiness, a complex genetic and environmental trait |
| title_full | HairNet2: deep learning to quantify cotton leaf hairiness, a complex genetic and environmental trait |
| title_fullStr | HairNet2: deep learning to quantify cotton leaf hairiness, a complex genetic and environmental trait |
| title_full_unstemmed | HairNet2: deep learning to quantify cotton leaf hairiness, a complex genetic and environmental trait |
| title_short | HairNet2: deep learning to quantify cotton leaf hairiness, a complex genetic and environmental trait |
| title_sort | hairnet2 deep learning to quantify cotton leaf hairiness a complex genetic and environmental trait |
| topic | Deep learning Neural network Machine learning Phenotyping Trichome Cotton |
| url | https://doi.org/10.1186/s13007-024-01149-8 |
| work_keys_str_mv | AT moshiurfarazi hairnet2deeplearningtoquantifycottonleafhairinessacomplexgeneticandenvironmentaltrait AT warrencconaty hairnet2deeplearningtoquantifycottonleafhairinessacomplexgeneticandenvironmentaltrait AT lucyegan hairnet2deeplearningtoquantifycottonleafhairinessacomplexgeneticandenvironmentaltrait AT susanpjthompson hairnet2deeplearningtoquantifycottonleafhairinessacomplexgeneticandenvironmentaltrait AT iainwwilson hairnet2deeplearningtoquantifycottonleafhairinessacomplexgeneticandenvironmentaltrait AT shimingliu hairnet2deeplearningtoquantifycottonleafhairinessacomplexgeneticandenvironmentaltrait AT warwicknstiller hairnet2deeplearningtoquantifycottonleafhairinessacomplexgeneticandenvironmentaltrait AT larspetersson hairnet2deeplearningtoquantifycottonleafhairinessacomplexgeneticandenvironmentaltrait AT vivienrolland hairnet2deeplearningtoquantifycottonleafhairinessacomplexgeneticandenvironmentaltrait |