Early detection of cancer is key for increasing treatment efficacy and improved survival rates. While there are diagnostic tests available for the early detection of a subset of cancers, these can involve invasive procedures which lowers their widespread use. Moreover, at present, these tools exist for a small minority of cancers; for too many, detection largely occurs at an advanced and incurable stage. In this project, Hani Goodarzi is harnessing the power of deep sequencing RNA to uncover tumor signatures though a simple blood sample. Goodarzi and his team have found that cancerous cells often aberrantly express non-coding RNAs which are not expressed by healthy cells due to the general transcriptomic dysregulation in cancer. These “orphan-noncoding RNAs” (oncRNAs) represent a biomarker for the presence of cancer cells, which continually secrete the oncRNAs into the bloodstream. Goodarzi and his team have designed a rigorous retrospective study in collaboration with clinical teams at UCSF; in this study, they will profile samples collected from 200 breast cancer patients before, during and after treatment. They will build a specialized machine learning platform that can efficiently and effectively mine cell-free RNA species in the resulting samples to perform risk assessment and observe changes in oncRNAs over the course of treatment. Finally, they will assess the generalizability of this platform to other cancer types. These studies will aid in developing assays which can lead to diagnoses earlier, less invasively, and for a broader array of cancers.