Data collection Guidelines for collecting and checking data Type of data z Quantitative - Height, diameter, density z Qualitative - Stem straightness Choosing traits for measurement and assessment • survival • dbh • height • stem volume • wood density, colour • timber strength, stiffness • timber defects • pulp yield • fibre length • stem straightness • axis persistence/forking • branch thickness • branch angle • pest and disease resistance • growth stress • tension wood • fodder production • fodder value • other traits? ? ? ? ? ? Choosing traits for measurement and assessment z Breeders aim to achieve genetic improvement in traits of economic importance z Breeders need to talk to the people (industry managers, farmers, etc.) who plant and use their tree species, to find out which traits are most important to the users z Examples : ¾ stem straightness is not very important for trees grown for pulpwood, but important for trees grown for sawlogs (bends in the stem reduce the recovery of sawn wood and therefore the value of the log) ¾ dry biomass/hectare, not volume/hectare is important for biomass energy users Selecting and breeding for a single trait, or for multiple traits z Breeding for a single trait is straightforward - we just rank the trees for the trait and choose the better trees for breeding and propagation z When breeding for two or more traits we must make “trade-offs” between traits. The tree with the largest stem volume may have very poor stem straightness - should we select this tree, if both traits are important to the user? Assessing traits z Objective or subjective scoring systems? z Objective - e.g. 1 = no flowering 2 = flowering z Subjective - e.g. stem straightness 1 = worst 2% of trees in trial 2 = next best 15% of trees in trial 3 = next best 33% of trees in trial 4 = next best 33% of trees in trial 5 = next best 15 % of trees in trial 6 = best 2 % of trees in trial Assessing stem straightness - subjective scoring system worst Prior to scoring, inspect trial and set proportions of scoring categories to approximate normal distribution - improves heritability of trait Stem straightness 1 2 43 5 6 best 33% 2% 15% Frequency Assessing traits z Best category gets highest score (gives consistency in constructing selection index) z An even number of categories (4, or 6) gives higher heritability than odd numbers of categories (3, 5, or 7) because we are forced to make decisions about the “average” trees - are they above or below the mean? 1 24 3 5 even odd 1 2 43 5 6 ⇐ ? ⇒ Axis persistence - objective scoring system 1 = stem axis forks at ground level 2 = stem forks in first quarter of tree height 3 = stem forks in second quarter of tree height 6 = axis persists to top of tree 4 = stem forks in third quarter of tree height 5 = stem forks in fourth quarter of tree height Forking defined as two or more leaders, stem diameter of smaller leader is more than 50% of diameter of larger leader just above fork Data collection z Indexing information on the field data sheets z Data sheets should be prepared with layout and treatment information included: replicate number, plot number, tree number, seedlot number, etc. [...]... 1 3 4 * * 1 1 4 4 * * 1 1 5 4 1.1 0 Check the data !!!!!!!!!!!!!!! There will always be mistakes in the data! Mistakes arise at different stages of the operation Read back the data from the computer screen, with somebody checking the field data sheet against the values which are being read out General tips for computer analysis of data Keep all the files for an experiment in one folder (directory) Check... replicates 5 rows and 12 columns spacing 3m between rows spacing 1.5m between trees within rows each seedlot occurs only once in any long column Assess the trial in field order !!! One line - one tree! One line on the data sheet should be used for each experimental unit (usually a tree) Measurements such as height and diameter are put in columns across the data sheet after the indexing columns Data sheet... Trees within a plot - same order for each measurement! 1…25 Tree number…… 1 2 3 4 6 7 8 9 11 12 13 14 16 17 18 19 21 22 23 24 5 10 15 20 25 Missing values - enter * Missing values are represented by the * symbol, for correct analysis by Genstat Blanks are not acceptable! If the value is 0 (for example the dbh of a tree which is 1.1 m high) enter 0, not * Missing trees and variates repl plot tree seedlot... Seedlot 1 1 1 4 1 1 2 4 2 5 7 2 2 5 8 2 ht03 ht04 dbh04 2.7 0.8 * 3.2 Collect data in field order • Indexing information should be in field order, NOT treatment order • Measure a replicate at a time, using same team, to avoid bias • Successive measures should be in the same order Latinised row column design for seedling seed orchard with 60 families Col 1 2 3 4 5 6 7 8 9 10 11 12 35 57 47 45... Keep a back-up copy of important files such as your original data file As you will most likely modify the original data file, work with a copy under a different name e.g benthamii2.xls Save your work frequently so it is not lost in a power failure, or if a program crashes Excel tips Edit\Goto\special\blanks - locate blank cells in the block of data you have entered =max(F2:F4000) - identify max value... Edit\Goto\special\blanks - locate blank cells in the block of data you have entered =max(F2:F4000) - identify max value in range F2:F4000 =min(F2:F4000) - identify minimum value =average(F2:F4000) calculate mean Data should make biological sense! Are these trees OK? height tree (m) diameter at ground level (cm) diameter at breast height (cm) 1 15.5 19.5 15.6 2 1.1 3.5 0 3 15.6 14.5 22.8 4 * * 13.0 5 1.1 4.2 2.1 . Data collection Guidelines for collecting and checking data Type of data z Quantitative - Height, diameter, density z Qualitative - Stem straightness Choosing traits for measurement and. important for biomass energy users Selecting and breeding for a single trait, or for multiple traits z Breeding for a single trait is straightforward - we just rank the trees for the trait and choose. straightness is not very important for trees grown for pulpwood, but important for trees grown for sawlogs (bends in the stem reduce the recovery of sawn wood and therefore the value of the log) ¾