[spatial & single cellomics] phase 1: Study on PDAC tumor microenvironment by single cell binding spatial transcriptome

*

6.19 On that day, I released the creation on gongzong 「 Space & Literature study group of single cell omics 」 's post , At the same time, several thinking problems have been put forward . Because the number of people is controlled , The final team members only identified 9 position , There are more than ten friends who contact me , I missed the bus , To you 「 sorry 」. If there are other groups in the future , I will inform you again . Study group course 、 video And so on and so on 「 Gongzong No. was released 」, You can pay attention to it .

*

Last Sunday , The group conducted The first 1 period Share , from TOP bacteria and Jeffery be the speaker , Respectively introduced ：

• 2020 Single cell binding spatial transcriptome study PDAC Tumor microenvironment
• science Recently published single cell space metabonomics technology

This tweet yes TOP Bacteria literature sharing Text version , video Explanation already in B standing Release ; The... Of this issue 2 Tweets yes Jeffery Literature sharing Text version , Reprinted from his gongzong No ： Student information programming self-study room

Let's go to the literature interpretation

background

This article was published in 2020 year 1 month , It is the earliest single cell binding idling article I can find .

The background of the article is mainly 3 spot ：

• scRNA-seq It will be dissociated before sequencing , Loss of spatial information
• Combining in situ hybridization and scRNA-seq It can solve the above problems to a certain extent , but ISH The limitations are also obvious , Only a small number of genes can be captured
• As early as 2016 In the year , Space transcriptome technology was born , But the limitation is the lack of single cell resolution

*

This article combines scRNA-seq And spatial transcriptome , The two technologies complement each other , To explore the tumor microenvironment of pancreatic ductal adenocarcinoma

*

result

1. Identify cell subpopulations

The technical route is relatively simple , The sample size of early articles is also small , The main content is based on the samples of two patients （ Paired samples are scarce ）

First, we made cluster annotation for single cell data of two patients , Then I took a look at A,B Consistency of annotation results of two patient subgroups .

Still by inference CNV The malignant cells were identified by the method of , Further from CNV It can be seen from the heat map A There are two groups of patients CNV Cells with obvious differences . It was confirmed by immunofluorescence A,B The presence of tumor cells in the patient .

2. Analysis of spatial transcriptome data

First of all, according to the HE Dyed pictures will A,B The patient's section is divided according to the tissue characteristics , From the spatial transcriptome data, we can also see some gene expression corresponding to the partition

Express data from idling only , Through the analysis process similar to single cell transcriptome , You can also see several groups spot, And it is consistent with the partition based on tissue slice

3.MIA( Multimodal intersection analysis )

In principle , Similar enrichment analysis , Is not complicated . The heat map on the right is viewed one by one , The deeper the red , It means that this cell type is in this region China and Vietnam are enriched . Through this analysis , Find out A Among patients , stay cancer region In addition to enrichment to tumor cells , It is also enriched in fibroblasts .

4. Analysis of cell subclasses

In this part, the ductal epithelial cells were analyzed in the same way 、 Macrophages 、DC cells . First, it is divided into sub categories , Reuse MIA Method to see the subclass in region The above enrichment .

5. stay cancer region in , Whether different tumor cells co locate with different other cell types

Take extra A Two sections of the patient , It is further subdivided in three slices cancer region. All three slices can be seen cancer cluster1 Co localization with fibroblasts .

6. From the point of view of cell state scRNA-seq And idling

utilize NMF From the single cell transcriptome data 3 Expression modules , The author focuses on stress-response Related modules . Then according to stress-response Module expression , take cancer region Of spot Divided into high and low groups , Continue the comparison between the two groups , find high This group of highly expressed genes . Obviously ,stress-response Related genes are already differential genes , Why do you do this ？ There are at least two purposes for this ：

• Increase the number of cells
• Reflect in addition stress-response Other features besides

The genes identified represent stress-response high Those of spot Regional characteristics of , And then use MIA Analysis stress-response high What cells will be enriched in the area of , Results show iCAF( Inflammatory fibroblasts ) Will be enriched in these areas .（ There is a similar conclusion above , The conclusion here is to further refine .）

TCGA Of bulk The sample also verified iCAF signature and stress-response module The relevance of

This result was also confirmed by immunofluorescence .

The article also used mismatched melanoma samples , It goes further MIA Applicability of analysis .

summary