单细胞测序数据FindAllMaker查找不同Cluster的Marker基因时报错

#快速找出所有细胞类型的差异/Marker基因

Rscript $scripts/seurat_FindAllMarkers.r --rds $workdir/05.cell_type_ann/pbmc.added.celltype.rds \

 -p FindAllMarkers --test.use  DESeq2   --logfc.threshold 0.25 ,执行此代码后报错:Warning: When testing 20 versus all:

every gene contains at least one zero, cannot compute log geometric means

Warning: When testing 21 versus all:

every gene contains at least one zero, cannot compute log geometric means

Warning: When testing 22 versus all:

every gene contains at least one zero, cannot compute log geometric means

Warning: When testing 23 versus all:

every gene contains at least one zero, cannot compute log geometric means

Warning: When testing 24 versus all:

every gene contains at least one zero, cannot compute log geometric means

data frame with 0 columns and 0 rows

Error in `group_by_prepare()`:

! Must group by variables found in `.data`.

• Column `cluster` is not found.

Backtrace:

    ▆

 1. ├─obj.markers %>% group_by(cluster) %>% ...

 2. ├─dplyr::top_n(., n = opt$top_n, wt = avg_log2FC)

 3. │ └─dplyr::filter(...)

 4. ├─dplyr::group_by(., cluster)

 5. └─dplyr:::group_by.data.frame(., cluster)

 6.   └─dplyr::group_by_prepare(.data, ..., .add = .add, caller_env = caller_env())

 7.     └─rlang::abort(c("Must group by variables found in `.data`.", glue("Column `{unknown}` is not found.")))

Execution halted

想问下是我输入文件的问题还是?我从第一部开始就是按照视频中的课程操作的?麻烦老师解答。谢谢老师


请先 登录 后评论

2 个回答

omicsgene - 生物信息
擅长:重测序,遗传进化,转录组,GWAS

可能是由于选择的差异分析方法和筛选条件太严格,导致没有找到差异的marker基因,你可以换个方法,并且降低筛选条件,例如:


Rscript $scripts/seurat_FindAllMarkers.r --rds $workdir/05.cell_type_ann/pbmc.added.celltype.rds \
 -p FindAllMarkers --test.use  wilcox   --logfc.threshold 0.2


例如这里有文献用MAST方法:


attachments-2024-05-weL7G6uq665561bf9889c.png

参考:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8104264/bin/NIHMS1638788-supplement-Supplemental_Publication_Material.pdf

在线阅读链接:https://doi.org/10.1161%2FCIRCULATIONAHA.120.048378

请先 登录 后评论


好的,谢谢谢谢老师解答。我再试一下。我总觉得0.25已经很小了,文献里面好多都是大于0.25的。

请先 登录 后评论