## 0/0 packages newly attached/loaded, see sessionInfo() for details.
## 0/0 packages newly attached/loaded, see sessionInfo() for details.
projectId = "53c53cd4-8127-4e12-bc7f-8fe1610a715c"
file_filter <- filters(
projectId = list(is = projectId),
fileFormat = list(is = "loom")
)
pfile = files(file_filter)
pfile$projectTitle[1]
## [1] "A Cellular Anatomy of the Normal Adult Human Prostate and Prostatic Urethra"
Very superficial filtering (to 60000 cells) and development of PCA
library(LoomExperiment)
f1 = import("/home/stvjc/.cache/R/hca/36e582f7c6e_36e582f7c6e.loom")
f1
names(colData(f1))
library(scater)
sf1 = as(f1, "SingleCellExperiment")
sf1
library(scuttle)
assay(sf1[1:4,1:4])
assayNames(sf1) = "counts"
litsf1 = sf1[,1:60000]
z = DelayedArray::rowSums(assay(litsf1))
mean(z==0)
todrop = which(z==0)
litsf2 = litsf1[-todrop,]
assay(litsf2)
litsf2 = logNormCounts(litsf2)
litsf2 = runPCA(litsf2)
Whatchaget:
if (!exists("litsf2")) load("litsf2.rda") # run code above, must have HDF5 in cache
metadata(litsf2)
## $last_modified
## [1] "20210616T032315.280827Z"
##
## $CreationDate
## [1] "20210616T030548.106180Z"
##
## $LOOM_SPEC_VERSION
## [1] "3.0.0"
##
## $expression_data_type
## [1] "exonic"
##
## $input_id
## [1] "5012cf4f-ba5c-4928-aea1-ab2891e34323"
##
## $input_id_metadata_field
## [1] "sequencing_process.provenance.document_id"
##
## $input_name
## [1] "D27PrTzF_Edn"
##
## $input_name_metadata_field
## [1] "sequencing_input.biomaterial_core.biomaterial_id"
##
## $optimus_output_schema_version
## [1] "1.0.0"
##
## $pipeline_version
## [1] "Optimus_v4.2.3"
> str(litsf2) # 22MB on disk (no quantifications)
Formal class 'DelayedMatrix' [package "DelayedArray"] with 1 slot
..@ seed:Formal class 'DelayedAperm' [package "DelayedArray"] with 2 slots
.. .. ..@ perm: int [1:2] 2 1
.. .. ..@ seed:Formal class 'DelayedSubset' [package "DelayedArray"] with 2 slots
.. .. .. .. ..@ index:List of 2
.. .. .. .. .. ..$ : int [1:60000] 1 2 3 4 5 6 7 8 9 10 ...
.. .. .. .. .. ..$ : int [1:23420] 13 20 22 23 31 33 34 35 36 37 ...
.. .. .. .. ..@ seed :Formal class 'HDF5ArraySeed' [package "HDF5Array"] with 7 slots
.. .. .. .. .. .. ..@ filepath : chr "/home/stvjc/.cache/R/hca/36e582f7c6e_36e582f7c6e.loom"
.. .. .. .. .. .. ..@ name : chr "/matrix"
.. .. .. .. .. .. ..@ as_sparse: logi FALSE
.. .. .. .. .. .. ..@ type : chr NA
.. .. .. .. .. .. ..@ dim : int [1:2] 382197 58347
.. .. .. .. .. .. ..@ chunkdim : int [1:2] 64 64
.. .. .. .. .. .. ..@ first_val: int 0
stvjc@stvjc-XPS-13-9300:~/CSAMA_HCA$ ls -tl /home/stvjc/.cache/R/hca/36e582f7c6e_36e582f7c6e.loom
-rw-rw-r-- 1 stvjc stvjc 1206062245 Jun 21 22:37 /home/stvjc/.cache/R/hca/36e582f7c6e_36e582f7c6e.loom
## 0/0 packages newly attached/loaded, see sessionInfo() for details.
## Object of class 'OlsSearch':
## query: smooth muscle
## requested: 100 (out of 47980)
## response(s): 0
Innovation here: walk through linked ontologies such as PR and present additional facets about the concept in focus
Limitation: the OBO representation in use is outdated and out-links are sparse
library(ontoProc)
cl = getCellOnto()
chk = ctmarks(cl)
Projects: - use rols to get more interesting information about terms into the app - update the ontology resources - go beyond OBO … but not all the way to OWL? Evaluate the UI/UX needed to broaden ontology usage - impacts: data integration, precision of annotation, cognitive efficiency