Excited to share our preprint introducing the first pipeline to identify Cell type Specific Intracellular (CSI) microbes from single cell RNA-seq data. Here's a tweetorial summary. biorxiv.org/content/10.110… 
Multiple recent studies have pointed to the functional importance of tumor microbiome. One eg., is the case of Fusobacterium in primary & metastatic colon tumors, which drives tumorigenesis, influences response to chemotherapy & binds to multiple human immune inhibitory receptors
Among many, the key challenges to systematically chart tumor microbiome are: 1. Remove reads due to contamination, 2. Whether microbes are intra/extra cellular and 3. if intra, which cell types within the tumor do they reside in?
Addressing the above three questions, our pipeline - *CSI Microbes* is the first to identify intracellular microbes that are cell-type-specific while filtering out contaminating microbes.
Our key idea is to identify microbial sequences from single-cell (sc) RNAseq datasets, such that the taxa containing those sequences are differentially abundant between distinct cell types.
We first validated CSI-microbes on a dataset of immune cells deliberately infected with intracellular bacteria Salmonella.
Next, we applied CSI microbes across five scRNA-seq smartseq2 datasets spanning 47 cancer patients across three cancer types (breast, melanoma and glioma).
We identify Streptomyces as differentially abundant in the tumor cells of 1 out of 9 breast tumor samples with enough cells tested.
We find evidence suggesting the existence of several intracellular bacteria and fungi in tumor cells in 2 out of 8 melanoma tumors.
No cell-type-specific bacteria were identified in our analysis of 12 brain tumors.
In sum, we present a computational approach and proof-of-concept for a pipeline to identify cell-type-specific intracellular bacteria using scRNA-seq.
