Nowadays the problem of early depression detection is one of the most important in the field of psychology. Social networks analysis is widely applied to address this problem. In this paper, we consider the task of automatic detection of depression signs from textual messages of Russian social network VKontakte users. We describe the preparation of users’ profiles dataset and propose psycholinguistic and stylistic markers of depression in text. We evaluate machine learning methods for detecting signs of depression from social media messages. The results of experiments show that psycholinguistic markers based features achieved 66% of F1-score on the binary classification task which is promising result in comparison with similar works. Copyright © 2019 for this paper by its authors.