For your analytical work with limited oilseed samples, it's generally better to determine protein content after oil extraction rather than before. This approach allows you to maximize the use of your sample while obtaining reliable protein values, though there are some important considerations to keep in mind.

Why protein determination after oil extraction works well: Research has shown that protein analysis can be successfully performed on defatted oilseed meal obtained through Soxhlet extraction. Studies on various oilseeds including chickpea, sunflower, sesame, and others have demonstrated that defatted samples actually provide more accurate protein measurements because the removal of oil eliminates interference during protein analysis. The Kjeldahl method, which is the standard for protein determination, works effectively on oil-extracted residues and often gives higher protein percentages on a dry weight basis since the oil fraction has been removed.

Temperature considerations and protein stability: Your concern about heat-induced protein denaturation during Soxhlet extraction is valid but manageable. Research indicates that moderate temperatures used in standard Soxhlet extraction (typically 60-70°C with hexane) cause minimal protein denaturation in most oilseeds. Studies have shown that significant protein denaturation in oilseed proteins typically occurs at temperatures above 75-85°C, which is higher than typical Soxhlet extraction conditions. For example, sunflower protein isolates showed denaturation temperatures ranging from 75-89°C, suggesting that standard oil extraction temperatures are below the critical denaturation range.

Accuracy of protein determination after extraction: The protein content obtained from defatted meal does represent the true protein content of the original seed. However, you'll need to calculate back to the whole seed basis using the formula: Seed protein % = (Defatted meal protein % × (100 - oil %)) / 100. This calculation accounts for the oil that was removed and gives you the actual protein percentage in the original seed. Studies have consistently shown good correlation between protein values calculated this way and direct measurements, making this approach scientifically sound for your research needs.

The sequential approach of oil extraction followed by protein analysis is not only practical for your sample limitations but also follows established protocols used in commercial oilseed analysis. This method will allow you to obtain both the oil you need and accurate protein data from the same sample, making efficient use of your limited material while maintaining analytical reliability.

I am not familiar with oilseed samples, but what about using NIRS for total Nitrogen in a first step, if you have very small sample size ?

Here in RG is this article from ethopia:

DOI 10.3389/fanim.2023.120344 :

Near-infrared reflectance spectroscopy using a portable instrument to measure the nutritive value of oilseed meals as livestock feed