(Popsci) - A new business called Replica Wines claims to produce "master forgeries" of well-known wines, deploying a throng of chemical instruments and a huge flavor database to blend near-identical versions of the wines from different grapes.

Ava Winery, meanwhile, takes a different approach, bypassing grapes entirely, and going straight for the molecules — combining flavor chemicals with ethanol and water to reproduce the experience of wine, without replicating the process.

It should be noted that it's a bit unfair to compare the two. Replica is part of an established company — the Colorado-based Integrated Beverages Group — with a line of commercial products and a team that includes a master sommelier and several distinguished winemakers. Ava is more of a lark, a thought experiment that might never have germinated outside of the hothouse conditions of San Francisco start-up culture. Its initial offering, a bottle of imitation 1992 Dom Perignon, is not yet for sale — and may never be. But these two very different companies showcase the ambitions, and limits, of chemical analysis when it comes to subjective qualities such as flavor.

According to Replica, its patent-pending proprietary method represents the "ultimate science and wine pairing," one that "takes the guess-work" out of making good wine. Working with Ellipse Analytics, an independent analytical chemical laboratory, the company claims to have assembled the world's largest database of alcoholic beverage flavor profiles — essentially, a database that pairs chemical markers with sensory effects.

To duplicate a wine, laboratory staff at Ellipse analyze and quantify its characteristic "macrocomponents," such as acidity, sugars, and tannins; and "microcomponents," volatile flavor chemicals such as linalool and methoxypyrazine. A panel of trained tasters also produces a subjective flavor profile, a record of perceived flavor notes and intensities. The goal is to align chemistry with sensory experience, using base wines from the company's own vineyards, as well as additional wines sourced from other California and Oregon winemakers, to make a blend that mimics the original. It's an iterative process, with analytic laboratory's results serving to guide the hands-on work of winemakers.

Once Replica's blend can effectively pass for the original in sensory tests, the laboratory then confirms the accuracy of the copy. "We ensure at least ninety percent chemical similarity to the wine by which each was inspired," promises Jaclyn Bowen, president of Ellipse Analytics.

Ninety percent sounds good, but given the exquisite sensitivity of the human sensorium, even a tiny difference, molecularly speaking, can matter quite a bit when it comes to actual flavor. So how does it go down? A tasting last week in the Popular Science offices in Manhattan placed Replica's Pickpocket side by side with The Prisoner, the popular and critically praised California red blend that "inspired" it. Although I'm not personally a huge fan of this style of red wine, the two were comparably big-boned, juicy, and smooth-drinking, and a similar deep purple color in the glass.