Not enough information to answer your question.

Thank you very much for your reply. I used spectrometer to measure treated cell and untreated as a control. I made four replicates. I got the value of treated cells in some samples higher than control. Therefore when I applied the equation, the value was negative. Now I need to use two way anova to compare between my samples. Normally I use Minitab software but unfortunately this software accept only positive values. So my question is how can I solve this problem. I repeated the experiment three times and got same negative values.

You have still not explained what you have measured (there are plenty of things that can come from  a spectrometer...what type of spectrometer is it: UV-VIS?  Are you looking at absorbance or wavelength or both?  Is this s single or dual beam instrument?) and the values obtained. What is 'the equation'?  There are many equations.  You state 'Now I need to use two way anova to compare between my samples"  Why?  How have you treated your cells?  For what reasons cannot you use the negative values as positives?  (This last question for you to understand and research what you are doing).

I am doing MTT assay

First, it's useful to try a little research before posting your question:  I recommend: http://bfy.tw/Bsvb

See also: https://www.researchgate.net/post/What_is_the_meaning_of_negative_value_percentage_of_inhibition_The_value_of_absorbance_of_sample_are_higher_than_the_absorbance_of_negative_control

Thank you

Even though your question is not very clear, often certain results are not as expected. For instance I have obtained negative residual variance before. We know theoretically variance should not be negative. I have gotten a negative R-squared before with Eviews and R-squared should not be negative. Any anomaly like that could arise out of a violation of a vital assumption or model inadequacy. It could also indicate that the non-negative statistic is actually zero.

Even you have negative values, it's ok, and u can use (graph prism pad), it is the best to use for this kind of data

Thank you Dhurgham for your reply,

What is graph prism pad, cold you explain more, please?

Thank you  Ette Etuk for reply,

Do you mean I can use zero instead of negative results?