The 2002 movie “All About Schmidt” struck an all-too-familiar chord with Dr. Monte Rouquette, Texas A&M AgriLife Research scientist, Overton, he said.
In the movie, Warren Schmidt, played by Jack Nicholson, is an insurance statistician who has taken great pride in his work of decades. Only days after his retirement, Schmidt returns to his workplace only to find years of his painstakingly collected data on paper files now stacked in boxes by a dumpster.
His young successor used a computer and just didn’t want to mess with paper files, Rouquette said. Schmidt’s data weren’t obsolete; his data storage method was.
This was a kind of epiphany for Rouquette, he said, as it echoed what he had often seen happen during his and other senior researchers careers: decades of research data lost because the newer generation of scientists found their recording methods outdated.
When he watched the movie, Rouquette was already working on a solution with Dr. Charles Long, resident director of research at the Texas A&M AgriLife Research and Extension Center at Overton, and Dr. Greg Clary, Texas A&M AgriLife Extension Service economist, Overton, now retired.
The solution was BeefSys, which today has become a massive archival database containing all the research work done on beef cattle, forages and pastures at the Overton center since 1967.
“This is not one experiment. This is many, many, many experiments,” Rouquette said. “There’s a lot of forage and animal performance data — a gold mine of data.”
Rouquette, himself, had research data dating back to times when data were entered on reel-to-reel tape, 5 1/4 inch floppy disks, computer punch cards and even handwritten journals.
For scientists who have been working in their disciplines before the advent of the personal computer, this is not an uncommon situation, he said.
However, Rouquette said, he finds the situation particularly worrisome today because there are fewer people doing the kind of multi-disciplinary work he and his colleagues have done for years.
“We’re going to do fewer grazing studies in the future than we have in the past,” he said. “Why is that? Because we have less funding and fewer scientists in forage research across the South. Why is that? Because forages and beef are not as commodity-identifiable as other commodities that may have a more basic or laboratory emphasis.”
That’s not to say that the work isn’t needed, Rouquette said. Environmental changes, such as drought patterns, new plant diseases and an evermore competitive market situation for beef and forage producers mean the research needs to go on.
“There is almost an urgent need for data archival systems for various crops, fertilizer and animal research,” he said. “Funding of all research is an issue, and scientists need access to previously conducted experiments to avoid having to reinvent the wheel.
“The stakeholder doesn’t recognize this redirection in forage-based and animal-production systems across the nation. He or she will still want information that can be applied to problems.”
Kelli Norman, AgriLife Research programmer analyst, Overton, explained that a relational database stores data together in tables, which have rows and columns much like a spreadsheet.
“But a database can have many different tables, depending on the scope of the data — the broader the scope, the more tables needed,” Norman said. “While all databases and spreadsheets allow the user to access records in a specific table, the strength of a relational database is the ability to access ‘related’ records across all tables, by use of a common field, referred to as a key.”
The key in BeefSys is assigning each animal a unique ID number, which allows researchers to access related data that may be spread across many tables: medical, calving, grazing, carcass traits, breeding and more, she said.
Norman has been managing the database since it’s inception in 2002, she said. Since then, it’s gone through several revisions, growing from a few records to include:
— Cow breed information, including F-1 Brahman X Herefords and F-1 Brahman X Angus crosses.
— Calving data for more than 1,000 cows, from 1967 to 2015.
— 650 cows with at least a five-year calving history.
— Birth-to-harvest data on 4,500 animals.
— 150 cow-calf trials with data for more than 5,000 cow-calf pairs, from 1980 to 2015.
— 120 stocker trials with data for more than 6,500 stocker calves, from 1967 to 2015.
— Feedlot data and carcass traits on more than 5,800 cattle, from 1986 to 2015.
More recently, she has added Brahman data from Dr. Ron Randel, an AgriLife Research scientist at Overton specializing in beef cattle reproduction. She also has added all the data that Dr. Charles Long collected while working with cattle at the Texas A&M AgriLife McGregor Research Center, before he became the resident director of research at the Overton Center.
Norman said the bull breeds include Angus, Brahman, Tuli, Simmental, Bonsmara, Limousine, Santa Gertrudis, Brown Swiss, Beefmaster, Senepol, Brangus, Hereford, Braunvieh, Charolais, Chianina, Holstein, Nelore and Romo.
Some of the specific animal data fields include:
— Calving, sire/dam information, birth date/weights and cow’s calving scores.
— Grazing site location/details.
— Stocking rates.
— Teat and udder scores at calving.
— Grazing method: continuous versus rotational.
— Stocking method: early versus deferred, and fixed versus variable.
— Supplemental feed, corn, cottonseed, fishmeal, soybean meal, feather meal.
— Weights, body condition scores, initiation dates of weaning.
— Average daily gains.
— Feedlot and carcass data, including feedlot average daily gains, in- and out- dates, dressing percentages, back fat, marbling scores and more.
Some of the specific pasture data fields include:
— Fertilizer application rates and times of application.
— Forage samples taken on test pastures at regular intervals.
— Pasture availability every 28 days.
— Nutritive quality every 14 days.
The data, Rouquette said, are not in summary form. Instead, entries are as raw numbers, direct from the original field records. The modern relational database format will allow researchers to compare performance of F-1 cows and their offspring on different pasture systems over an array of widely varying climatic conditions.
He foresees scientists using it for statistical analysis, creating decision-making tools for stakeholders in such areas as sustainable pasture management strategies, economic assessments and risk studies –both biological and financial – associated with pasture livestock operations.