I am trying to find the Mellin transform of $\sin x $, put in other words to solve:
$$\int^{\infty}_0 x^{s-1}\sin x \mathrm{d} x $$
And I know that the answer is:
$$\Gamma(s) \sin \left(\frac{\pi s}{2}\right)$$
From several tables on the internet but I couldn't find any justification.
How can this identity be proven?
The basic idea is to use Euler's formula: $\sin x = \dfrac{e^{ix}-e^{-ix}}{2i}$.
Let's look at the Mellin transform of $e^{ix}$:
$$\int_0^\infty e^{ix} x^{s-1}\,dx = \begin{bmatrix} x=it \\ dx = i\,dt\end{bmatrix} = \int_{0}^{-i\cdot\infty} e^{-t} (it)^{s-1} i\,dt = -i^s \int_{-i\cdot \infty}^0 e^{-t}t^{s-1}\,dt.$$
For certain values of $s$ (I will leave it to you to work out the details), you can deform the integral over the negative imaginary axis to an integral over the positive real axis: $$\int_{-i\cdot \infty}^0 e^{-t}t^{s-1}\,dt = -\int_0^\infty e^{-t}t^{s-1}\,dt = -\Gamma(s).$$
(Add a large quarter circle in the fourth quadrant, use Cauchys integral theorem and estimate $f(z) = e^{-z}z^{s-1}$ on the new quarter circle.)
Summing up, we get that Mellin transform of $e^{ix}$ is $$i^s\Gamma(s) = \exp(i\pi s/2) \Gamma(s).$$
Similarly, the Mellin transform of $e^{-ix}$ turns out to be $$i^{-s}\Gamma(s) = \exp(-i\pi s/2) \Gamma(s).$$
Forming the appropriate linear combination, the Mellin transform of $\sin x$ ends up as $$ \frac{\exp(i\pi s/2) + \exp(-i\pi s/2)}{2i} \Gamma(s) = \sin\frac{\pi s}2 \Gamma(s).$$