CAVS—Novel in silico selection strategy of specific STAT inhibitory compounds

• Specific STAT – targeting compounds of the phosphotyrosine (pTyr) SH2 interaction area.
• CAVS – five step comparative virtual screening tool.
• Selection criteria to identify specific STAT inhibitors.
• Developing of a general – purpose pipeline for different experiments and protein families in the future.

Signal transducers and activators of transcription (STATs) are a family of proteins activated by different stimulating factors, including interferons, interleukins, growth factors and oncoproteins. Their role in many diseases has been proven, so there is a strong demand to find specific strategies for STAT inhibition. Our searches for specific STAT-targeting compounds focused on exploring the phosphotyrosine (pTyr) SH2 interaction area. Current selection strategies are insufficient, thus, we developed a new pipeline strategy. In this work we used our recently built 3D models for all human (h)STATs (1–4, 5A, 5B and 6). To select specific inhibitors for the STAT protein of interest, we designed and implemented a five step comparative virtual screening tool, which we named – CAVS (comparative approach for virtual screening). CAVS introduces the ‘comparative binding affinity value’ (CBAV) and ‘ligand binding pose variation’ (LBPV) as selection criteria to identify specific inhibitors of STATs. In a five-step approach, including pre-screen, primary filtering of inhibitors, re-screen, secondary filtering of inhibitors and graphical inspection and final selection, CAVS leads to selection of specific STAT inhibitory compounds. CAVS was tested on a small ligand library of 130 000 natural products and 5.7 million of clean leads for the selection of STAT1 or STAT3-specific inhibitors. With the use of a designed set of Python scripts for data managing and filtering, CAVS allows to convert the comparative virtual screening procedure into an automatic pipeline and to effectively analyze virtual screening results from standard Surflex-Dock 2.6 output files. We also adapted CAVS as a general-purpose pipeline, which will allow adapting written code to different experiments and protein families in the future.