Most of our current knowledge on planet formation is still based on the analysis of main-sequence, solar-type stars. Conversely, detailed chemical studies of large samples of M-dwarf planet hosts are still missing. We aim to test whether the correlations between the metallicity, individual chemical abundances, and mass of the star and the presence of different type of planets found for FGK stars still holds for the less massive M dwarf stars. Methods to determine in a consistent way stellar abundances of M dwarfs from high-resolution optical spectra are still missing. The present work is a first attempt to fill this gap. We analyse in a coherent and homogeneous way a large sample of M dwarfs with and without known planetary companions. We develop for the first time a methodology to determine stellar abundances of elements others than iron for M dwarf stars from high-resolution, optical spectra. Our methodology is based on the use of principal component analysis and sparse Bayesian's methods. We made use of a set of M dwarfs orbiting around an FGK primary with known abundances to train our methods. We applied our methods to derive stellar metallicities and abundances of a large sample of M dwarfs observed within the framework of current radial velocity surveys. We then used a sample of nearby FGK stars to cross-validate our technique by comparing the derived abundance trends in the M dwarf sample with those found on the FGK stars. The metallicity distribution of the different subsamples shows that M dwarfs hosting giant planets show a planet-metallicity correlation as well as a correlation with the stellar mass. M dwarfs hosting low-mass planets do not seem to follow the planet-metallicity correlation. We also found that the frequency of low-mass planets does not depend on the mass of the stellar host. These results seem in agreement with previous works. However, we note that for giant planet hosts our metallicities predict a weaker planet metallicity correlation but a stronger mass-dependency than photometric values. We show, for the first time, that there seems to be no differences in the abundance distribution of elements different from iron between M dwarfs with and without known planets. Our data shows that low-mass stars with planets follow the same metallicity, mass, and abundance trends than their FGK counterparts, which are usually explained within the framework of core-accretion models.

Cone search capability for table J/A+A/644/A68/tablea1 (Basic properties of the stars analysed in this work)