استفاده از مدل رگرسیونی گام به گام جهت شناسایی صفات گیاهی مرتبط با عملکرد در برنج (oryza sativa) در شرایط تنش خشکی

در برنامه‌های به‌نژادی و اصلاح نباتات، برآورد میزان عملکرد بر اساس تغییر در صفات موثر گیاهی از اهمیت بالایی برخوردار است که با استفاده از مدل‌سازی رگرسیونی می‌توان آن‌ها را شناسایی کرد. به همین منظور 124 لاین از نسل نهم حاصل از تلاقی دو رقم برنج (اهلمی طارم و درفک) را در یک آزمایش در قالب طرح بلوک‌های کامل تصادفی با سه تکرار در دو سال زراعی 96-1395 در دانشگاه گنبد کاووس، مورد بررسی قرار گرفتند. آبیاری تا مرحله حداکثر پنجه‌زنی در شرایط بدون تنش (غرقاب) انجام شد. از این مرحله تا انتهای دوره رشد آبیاری به طور کامل قطع گردید. با استفاده از رگرسیون گام به گام، از بین صفات اندازه‌گیری شده، چهار صفت وزن بوته، وزن خوشه، وزن صد دانه و شاخص برداشت که در افزایش عملکرد بیشترین نقش را داشتند، شناسایی شدند و با توجه به همبستگی منفی موجود بین دو متغیر شاخص برداشت و وزن بوته، فرضیاتی مطرح شد که در هر فرض مقدار افزایش عملکرد برآورد گردید. نتایج نشان داد در صورتی که همبستگی بین صفات شاخص برداشت و وزن بوته قابل شکستن نباشد، عملکرد تیپ ایده‌آل 197/61 کیلوگرم در هکتار افزایش خواهد یافت. در صورتی که با افزایش وزن بوته و شاخص برداشت در حد متوسط بماند، عملکرد تیپ ایده‌آل 314/64 کیلوگرم در هکتار افزایش می‌یابد و چنانچه همبستگی بین وزن بوته و شاخص برداشت شکسته شود، مقدار افزایش عملکرد 697/52 کیلوگرم در هکتار خواهد بود. روش مورد استفاده در این تحقیق به علت اینکه به اختلافات ژنتیکی بین لاین‌ها توجه دارد، می‌تواند راهگشای متخصصان اصلاح نباتات برای انتخاب صفات موثر بر عملکرد برنج باشد.

using a step-by-step regression model to identify plant traits related to yield in rice (oryza sativa) under drought stress conditions

introduction: in order to select their breeding goals among different physiological traits, plant breeders need to categorize the limitations as well as the capabilities of plants. this issue has led to the emergence of a concept called ideal type. achieving the ideal type of agricultural plants requires the use of appropriate statistical methods. the aim of this study is to introduce the method of using regression modeling to determine the ideal type of crop plants on the rice plant.in breeding programs, it is very important to estimate the amount of yield based on the change in effective plant traits, which can be identified using regression modeling. in plant breeding, correlation between traits is of special importance, because it determines the amount and type of relationship between two or more traits. in plant breeding, correlation between traits is of particular importance, because it determines the degree and type of relationship between two or more traits. correlation between different traits can help researchers in indirect selection using traits that are easier to measure in order to achieve self-sufficiency.materials and methods: for this purpose, a study on 124 lines of the ninth generation of two varieties of rice (ahlomi tarom and dorfak) was conducted at gonbad kavus university, iran, using a randomized complete block design with three replications in two growing seasons of 2015 and 2016. irrigation was done until the maximum stage of tillering in stress-free conditions (flooding). from this stage until the end of the growth period, irrigation was completely stopped. among the measured traits, using stepwise regression, four traits of plant weight, panicle weight, 100-seed weight, and harvest index were identified that had the greatest role in increasing yield. then the correlation between the selected traits and the rate of increase in performance was evaluated and the rate of increase in performance (percentage) resulting from that trait in relation to the total increase in performance was calculated. regarding negative correlation between the harvest index and plant weight, three hypotheses were evaluated, that in each assumption the amount of increase in yield was estimated. data analysis was carried out using sas software.results and discussion: considering the existing negative correlation between the two variables of harvest index and plant weight, assumptions were made to determine the ideal type, the results indicated that if the correlation between the harvest index and plant weight, is not breakable, ideotype yield variation would have an increasing of 197.61 kg/ha. if with increasing plant weight, harvest index stay at moderate level, it would be an increasing of 314.64kg/ha and if correlation between plant weight and harvest index is breakable, it would be an increasing of 697.52 kg/ha. the method used in this study, due to the fact that the genetic differences between the lines are noticeable, can be a way for the breeders to move towards yield increasing in rice cultivars.conclusion: the main purpose of this article is to introduce the method of using regression modeling in determining the ideal type of crops. in this study, the ideal type of rice was determined using regression modeling. and the four characteristics of plant weight, cluster weight, 100 seed weight and harvest index had the greatest role in increasing the yield. considering the existing negative correlation between the two variables of harvest index and plant weight, assumptions were made to determine the ideal type. the results showed that if the relationship and correlation existing between some traits undergo changes, it can be used for the benefit of performance. if the main goal of the research is to determine the optimal type of rice for the gonbad region, it is better to study and research more genotypes in a few years.