Submitted to: Rice Technical Working Group Meeting Proceedings
Publication Type: Proceedings
Publication Acceptance Date: January 8, 2010
Publication Date: February 22, 2010
Citation: Hancock, T.A., McClung, A.M., McCouch, S.R., Eizenga, G.C. 2010. Phenotypic data collection and management using barcodes, MSExcel and MSAccess. In: Proceedings of the 33rd Rice Technical Working Group Meeting Proceedings, February 22-25, 2010, Biloxi, MS. CDROM Technical Abstract: A “Rice Diversity Panel” composed of 409 purified Oryza sativa accessions originating from 79 countries was developed in order to conduct an association mapping study. The methods used to collect and manage the phenotypic data on plant morphology, seed morphology, grain quality, and selected agronomic traits for the association mapping are described in this study. To better understand how plant and seed traits were selected during domestication, 133 accessions of the rice (O. sativa) ancestral species, O. rufipogon were included in this phenotypic evaluation with the objective of including them in the association mapping. The other long-range goal is to explore the genetic basis of transgressive variation that is observed after hybridization both within O. sativa and between O. sativa and O. rufipogon. Trait data for over 30 plant, seed, and grain quality characteristics were collected on the 409 rice accessions. The accessions were evaluated in a randomized complete block design conducted in the field at Stuttgart, Arkansas with two replications during two different years. Three representative plants were chosen from each replication for evaluation, thus phenotypic data were collected from twelve plants of each accession. In a similar study, 133 O. rufipogon accessions were grown in the greenhouse during two different years. Data were collected for 24 plant and seed traits from three plants of each O. rufipogon accession each year totaling six plants for each accession. These two studies produced over 165,000 data points that needed to be organized by accession, year, replication, and trait for accurate analysis. Microsoft Excel and barcodes were initially used to assist in the collection and organization of the data. As the dataset grew, we began using Microsoft Access to quickly look at averages by accessions, by one or more years, as well as by multiple traits. Unique identifiers were created for each plant in the study. Tags which contained barcodes with these unique identifiers were placed with each plant in the field or greenhouse. This was particularly important for some traits where there was wide variability among the plants for a given accession (e.g., days to heading 50 to >120 days). Envelopes for panicle harvest were labeled with the unique barcode identifier along with categorical ratings for eight phenotypic traits being recorded during harvest (shattering, leaf pubescence, plant type, awn presence, awn size category, panicle type, lodging incidence, plant height, and plant number in the row). All possible ratings for the traits were listed on the envelope, thus one only needed to circle the appropriate score. These “envelope notes” were transformed into electronic format using a Symbol barcode scanner with TracerPlus software while data on the single panicle in the envelope (panicle length, number of primary branches, number of whole seeds per panicle, number of unfilled florets per panicle) also were recorded electronically. To harvest seed from the individual field grown O. sativa plants, tags were removed from the stake, threaded through a plastic zip-tie, and wrapped around the plant tillers. The tillers were cut with a sickle, seed threshed using a bundle thresher, and seed collected in an envelope labeled with the unique identifier barcode and other pertinent information. Tags were placed in the envelope to aid in quality control. Traits were recorded on hulled and dehulled seed using an image analysis system. For recording the data, a subsample of seed from the plant harvest was placed in a coin envelope labeled with the unique barcode. Later, this sample was placed on the scanner and the WinSEEDLE software counted the seed and measured the length, width, volume and surface area of the seed. Afterwards, the sample was dehulled and these traits were measured again on the dehulled seed. Subsequently, the dehulled sample was prepared for quality analyses (alkali spreading value, amylose and protein content) using the appropriate methods, keeping the unique identifier with the sample throughout these analyses. In conclusion, using Microsoft Excel and Access allowed us to easily and quickly manage a large amount of phenotypic data that facilitated quality control and analysis. The methods of phenotypic data collection and management used in this study are applicable to collecting/managing data for large breeding programs and germplasm collections.