"Homologous recombination is a mechanism by which DNA double-strand breaks are faithfully repaired using the sequence information present on an intact duplex DNA. The efficiency of this mechanism is crucial for cell survival, and its accuracy is of prime importance for genomic stability. Although being the signature reaction of homologous recombination, the search for homology performed by the broken molecule coated with the Rad51 recombinase (called the “nucleoprotein filament”) is poorly understood: how a single entity is able to scan the megabases of the genome to rapidly find its unique homologous target remains a mystery. My post-doctoral project aims at tackling this long-standing ""needle-in-a-haystack"" problem. For this, I introduce new concepts and approaches to elucidate how this potent search takes place in a massive excess of competitor DNA.First, we hypothesize that several sections of the nucleoprotein filament are able to sample different duplex targets at the same time. This multiplexed search process by the nucleoprotein filament will be studied in vitro and in vivo. In particular, I will develop a technique derived from chromosome conformation capture to isolate and quantify “multiple invasions” performed by the nucleoprotein filament in vivo.Second, I intend to identify the functional substructure of the nucleoprotein filament, which is not a simple homogeneous assembly of Rad51 along the single-stranded DNA. Instead, it embeds other proteins, like the translocase Rad54 and the Rad51 paralogs, which exert a positive role on homology search through still unknown mechanisms. In particular, I will study how these proteins participate in the rapid rejection of pairing to heterologous molecules.Third, I will study the processing of the “multiple invasions” intermediate, and the potential threat it represents for genomic stability.This study will provide unprecedented insights into the key mechanism of homology search during homologous recombination."