NSIP for Romneys

Contents:

Talk given to Romney Dinner at MD Sheep and Wool Festival

Tips for using the NSIP Excel Workbook

Talk given to Romney Dinner at MD Sheep and Wool Festival 4/30/04

Performance Recording System (PRS) concept: to find which genetic lines improve the most upon breed-specific expectations for the values of meaningful, objectively measurable, heritable, variable traits. Good example: 60-day weight (meets all criteria); bad example : “eye appeal” – meaningful, heritable, variable, but not objectively measurable; bad example: “has two ears” – meaningful but little variability.

Although ewes contribute 50% or more to anything that is heritable, all the statistical methods focus mainly on rams. Rams have more progeny per animal, which gives greater precision in estimating differences from expectations.

Within-flock PRS: this can be do-it-yourself or use-a-(proprietary) system like Studfax (New Zealand) or NSIP (USA) or Lambplan (Australia). Example: you could see which of your three rams, each of whom bred more than 10 ewes, had the best 60-day weights for his lambs. If you already do this, you probably use published (in the SID manual) NSIP adjustment coefficients to adjust for different distributions among the progeny of the three rams in birth type, maternal age and rearing type. (The coefficients, by the way, are built from Suffolk data.) You might account also for artificial feeding, health problems, weaning groups, time of year born. You could never tell, however, which ram would have the best 60-day weight for progeny in a very different environment than yours.
Cross-flock (“Inter-flock”) PRSs: Lambplan (Australia); SIL (NZ); NSIP (USA). These systems track the progeny of rams either directly as breeding rams move from one farm’s ewes to another’s or indirectly through the progeny of the sons and daughters of those rams breeding on farms at, and away from, “home.” By recording how lines in varying locales perform (e.g. TN lamb-on-pasture and also NY winter barn-lambing), a cross-flock PRS spots the genetic lines that most consistently exceed expectations. It also tells you how precise those estimated deviations are. This allows potential buyers to weigh those deviations without sentiment, prestige or word-of-mouth. For example, prudence dictates you should buy a ram who’s had 100 lambs averaging 58 pounds at 60 days rather than one who’s had one lamb that weighed 68 pounds at 60 days. If you are bold venturer, though, you might still go for the second one.

The basic statistic of NSIP is the Expected Progeny Difference [EPD, = half the Estimated Breeding Value]. The units of EPD are the units of the characteristic being evaluated; for example, the EPD for fleece weight on a ram might be + 1.1 pounds. PROGENY of this ram are expected to have fleece weight that exceeds average by 1.1 pounds. Example: a ewe with an EPD for lambs/100 lambings of + 20; her PROGENY are expected to have 20 more lambs in 100 lambings than the progeny of comparison ewes.

Why should Romney breeders get into NSIP? The best reason is to work together on improvement. NSIP can use inputs from small and large flocks, high and low profile, show jockeys and spinning flocks and commercial. The highest initial payoff of NSIP is to breeders, but in the long term the highest payoff is to the breed itself.

To take part in NSIP you need to record systematically certain data that you already record (e.g. birth weight and type) and some that you probably get less systematically. For example, the program needs a value for 60-day weight on every lamb surviving at your place, not just your best ones. You need to write those data onto entry forms, either paper-only or (vastly preferred) electronic. The electronic form is an Excel workbook. Your data are pooled with those of all other Romney participants and run on the big computer at Virginia Tech once a year.

A common beef with NSIP is this once-a-year run, but remember that statistically valid cross-flock estimates require several years worth of data. We need a long perspective. Which genetic lines to use from within or to buy from without for breeding in the fall 2008 will use EPDs from the 2007 run, and to have good precise cross-flock EPDs sometime in 2007 we need to have data from at least several thousand lambings gathered from at least several dozen farms sent in properly in 2004-5-6 and 7. That requires we get going this year at good speed. [This was said in 2004; since Roomney NSIP startup in 2004 involved few breeders, make everything to a year later.]
I have not discussed the nuts and bolts of data collection and entry here; there isn’t time. I would be glad to meet with anyone this weekend to give more details. Please also see my article on the ARBA web site, a redo of one published in the Romney Ramblings of fall 2003.

The rest of this text was not spoken at the dinner at Maryland, but added since as I think over our first year’s experience and comments I’ve received.

Not many people have joined the Romney NSIP group so far. Reasons for the low participation breed-wide include (1) general dislike of record-keeping (2) technical problems with the measurements (“I don’t own a scale”) (3) lack of time to prepare the observations systematically (4) belief that the traits recorded are not important even for meat-lamb breeding (5) “my flock is too small” (6) recognition that some of the traits evaluated (e.g. fertility) are irrelevant to breeding show sheep. Each one of these reasons will keep someone out. Reason 1 is the lamest, reasons 5 and 6 the hardest to overcome.

Everything I have said and written about the ability of NSIP to improve one’s own flock and the breed in general hinges not just on careful cooperative data collection but also on breeders’ making decisions that use NSIP findings. That must be crystal-clear for us to benefit as a breed from NSIP. The most attentive recording, the most powerful statistical analysis, will not take us anywhere if we don’t use the findings to improve our sheep.

Much as I would like to set aside hurdles for participants, I have to ask that all data come to me already entered into the NSIP Excel workbook by you. I will be glad to send you the workbook electronically. It does take a lot of time to turn existing flock records into NSIP-ready format. I can’t promise to do it for anyone else though I cautiously made such an offer in the past (no one took me up on it). Equally important for me as coordinator is the power of e-mail to communicate with NSIPers both actual and potential. I really can’t do U.S. mail too. Going all-electronic may cost us a couple of participants whose loss I will regret.

Also on this site is a cover letter I have sent with the Excel workbook, meant to help Excel beginners.

NSIP Excel Workbook Tips

Dear Romney Breeder,

Here is the Excel workbook, NSIP version 3.1.1, into which you will record your data for NSIP. [Editor's note: the workbook is not actually linked to this page.] I’m afraid that printing out the data entry format has some people flummoxed. As an Excel beginner, I give some suggestions below that should help. Excel mavens can skip the next paragraph.

Download the workbook to your hard disk. Make a backup copy right away. The workbook, which you may call a spreadsheet, is in four sections. Two are purely instructional, two are for entering your findings. Print “Instructions” (section 1) entirely and “Codes” (Section 4) entirely.

The actual data entry forms are called “Mating-Lambing-Weaning” (section 2) and “Postweaning Weight and Wool” (Section 3). DO NOT TRY to print these in their entirety. To make paper versions for pencil use, follow these steps: (a) OPEN section 2 (M-L-W) (b) SELECT FOR PRINTING rows 8-47 and columns A through BA. This gives a table 40 rows by 53 columns, spread over three pages (c) check the preview. You should see three pages, each with column headers and 38 rows for data entry (d) print the three pages (e) now open section 3 (P-W-W), select rows 8-47 again but select only columns A-V. Preview (this is a single page) and print this table of 40 rows by 22 columns. Make photocopies .

Before you enter anything, go to www.nsip.org and download the PDF files for terms and frequently-asked questions. Please email me with questions not answered at those sources; on those I can’t answer right off I’ll dig deeper on for you. Refer also to the article I wrote about NSIP on the ARBA website www.americanromney.org

Most people joining in 2004 will be sending in data on lambs born 2004; if you have complete data for 2003, talk to me soon. We propose to send in those 2003 findings around July 1 2004, but that date is flexible. If you want me to wait a little longer for you, talk to me soon. It can be done.

A couple of Excel or Internet issues: (1) Go ahead and “enable” the macros to open. All NSIP staff and I as Romney data coordinator are using up-to-date protection. Be sure you are, too.

Excel for Dummies or another manual is a great help for an Excel beginner.

Be very careful if you decide to print your data. This is probably a good way off for most people. BE SURE to select in advance the exact rows and columns you want printed, or else the entire workbook (70+ pages of blank cells ) will print.

NSIP issues: The NSIP data base, when it looks at ewes, looks at the outcome of every conception it can track. Every line is the outcome of a recognized conception going to term or near-term. Thus a ewe may have three lines in one year if she has triplets. An NSIP macro makes it quick to reproduce the ewe-identifying and sire-identifying columns for those triplets in the example, but you must individualize the outcomes.

Liveborn lambs are identified at first by their flock eartag. Some breeders assign a flock number also to every recognized aborted fetus and to every stillbirth. If you haven’t done this before, consider starting to do it. An alternative for NSIP record-keeping is to record for each recognized conception that never took a breath just the birth date and the Livability code. You make the decision whether a lamb born dead at (say) 140 days estimated since conception is to be called an “abortion” (Livability code ‘1’) or a “stillbirth” (Livability code ‘2’). You record the code in the appropriate column (Lamb LIV) and enter a weight if you took it. The program will credit the ewe with a conception when it sees the birth date and the livability code.

If a ewe was exposed but you never ever saw any evidence of conception, leave all lamb data cells blank. Ewes that were not exposed at all should not enter the data base; these would be, for example, most lambs and also some aged ewes kept for old times’ sake.

Rams are done differently, entered on a single line so they can get a unique identifier.
NSIP does not charge for recording rams.

For example, if you have 50 ewes, 4 breeding rams and 80 recognized conceptions with at least one from each ewe you’ll enter 84 lines of data, but will be charged for 50 breeding animals. If , of those same 50 ewes, 49 got pregnant and had 80 recognized conceptions with one ewe being open, you would still have 84 lines of data but NSIP should charge you only for 49 breeding animals.

(1) I assume that all breeding animals resident in North America will be registered with ARBA. If you believe you are raising purebred Romneys but some of your breeding animals are not registered with ARBA, please talk to me, the Romney coordinator for NSIP. We could still use your data. I hope there is no one in North America in this situation, but if there were we would not want that person unable to take part. A breed society registration number is not actually required for identification, but animals without one will be special cases. If, as I hope will happen, South American Romney breeders want to join NSIP they can use their breed society numbers.

(2) If your breeding sheep has a black factor but looks white the animal should have letters in its registration number like bw or BW. If it looks natural-colored and has at least one natural-colored or black-factor parent, it should have BB or bb in its registration number. If it looks natural-colored and neither parent has a BW or b or BB or B in its registration number, that’s interesting. Hmmmm. Either the letter got dropped somewhere by mistake or you have a new mutation or a very-long- hidden color factor. For Romneys in NSIP, let’s set the following custom: If there is either a lower-case (b) or an upper-case letter B in anywhere in the registration number, regardless of the sheep’s appearance, enter a capital B in the OPT column. This will allow NSIP to adjust for presence of a black factor if we decide that is necessary.

(3) The only traits on which NSIP can do EDPs for us now are those on the data sheet either explicitly (e.g. lamb’s 60-day weight or conceptions per ewe per year or mid-side average fiber diameter) or by recalculation (e.g. maternal milk production, which captures the maternal effect on preweaning weight. When we get enough Romneys into NSIP there will be a chance for us to pick one or more traits of special importance to us, e.g. a scale of fleece uniformity. We aren’t there yet. Tell me what you would like to see NSIP do, but don’t demand it now.

(4) The column for codon 171 in the latest intake form I personally recommend be left blank. Very few people have data on Romneys in the U.S. I do not want it to look now as if NSIP believes testing at codon 171 is worth doing in Romneys now in the same way as (say) maternal milk production is. If we as a breed decide it is, or commit to another codon or to another indicator of resistance or susceptibility to scrapie, we can start entering those data. We might very well do so, but since we haven’t yet, I don’t want NSIP to be seen as involved with the decision.

(5) Your flock number for NSIP is made by a combination of breed, state and order within the state. I will assign it.

Anchorage Farm Registered Romneys 8 Mynderse Street Saugerties NY 12477-1704 ...in the mid-Hudson Valley
Anchorage 280-03 (AI)	Graeme L. Stewart, Manager 845-246-7685 Anchorage1@earthlink.net Steve and Elizabeth Shafer, Owners 845-246-4947 212-222-5703 sqs1@columbia.edu shpcount@earthlink.net

Home     |       History of Anchorage Farm    |    NSIP Info    |    OPP Statement