Dear users,

Instead of counts (non-negative integers) required by DeSeq2, I have negative numbers and fractions,

and I need to use DeSeq's statistics to find out significantly up/down genes.

Since for my data, min=-10 and max=10, I can convert it into counts by 

1. first multiplying with a constant eg. 1000 to preserve the decimal precision.
2. and then add some constant value eg. 10000 to make everything positive.

Would it hamper my analysis?

You can't use negative numbers with DESeq2. The model only makes sense if the input is counts or estimated counts of reads or fragments.