The clonalCluster() function cannot generate clusters. #433
Comments
|
Hey @HaiqiaoSun Thanks for reaching out. Please make a reproducible example for this issue so I know what you are experiencing. Also with the example, please make sure to include the output of sessionInfo(). Thanks a lot, |
|
Thank you for your prompt response. Here is the data and example I used. https://gist.github.com/HaiqiaoSun/698bc2d3e14950feb085fa19a6d56077 > library(scRepertoire)
> csv1 <- read.csv("E:/R/filtered_contig_annotations.csv", stringsAsFactors = F)
> combined <- combineBCR(csv1, samples = c("sp_exp"),
+ ID = c(NULL))
> head(combined[[1]])
barcode sample IGH cdr3_aa1
1 sp_exp_ATCCACGCCGCAGTCCAGTA sp_exp IGHV3-33.IGHD3-3.IGHJ6.IGHM CARDRQVLRFLEWLLPYYYYGMDVW
2 sp_exp_AGTATCTAACCACTTGTAAG sp_exp IGHV3-7.NA.IGHJ4.IGHM CARDGGDDGYW
3 sp_exp_AACTGGCGAGCAATAACACT sp_exp IGHV3-21.IGHD3-10.IGHJ4.IGHM CARIPPVTMVRGVIKFFDYW
4 sp_exp_GACAGCGAATGGAGTTAAGG sp_exp IGHV3-53.NA.IGHJ4.IGHM CARDSYSSHYW
5 sp_exp_TTGTCGAAGGCTCATACGCT sp_exp IGHV4-34.NA.IGHJ6.IGHM CASGIKQQLAYYYYGMDVW
6 sp_exp_TAGCGTTCGTAACGATTCAA sp_exp <NA> <NA>
cdr3_nt1 IGLC cdr3_aa2
1 TGTGCGAGAGATCGTCAAGTATTACGATTTTTGGAGTGGTTATTACCCTACTACTACTACGGTATGGACGTCTGG IGKV2-24.IGKJ2.IGKC CTQATQSPYTF
2 TGTGCGAGAGATGGCGGTGACGACGGGTACTGG IGKV2-24.IGKJ2.IGKC CTQATQFPYTF
3 TGTGCGAGAATACCTCCGGTTACTATGGTTCGGGGAGTTATTAAATTCTTTGACTACTGG IGKV2-24.IGKJ2.IGKC CTQATQFPYTF
4 TGTGCGAGAGATTCGTATAGCAGCCATTACTGG IGKV2-24.IGKJ1.IGKC CTQATQFPWTF
5 TGTGCGAGTGGAATTAAGCAGCAGCTGGCCTACTACTACTACGGTATGGACGTCTGG IGKV2-24.IGKJ4.IGKC CTQATQFPHTF
6 <NA> IGKV2-24.IGKJ5.IGKC CTQATQFPHF
cdr3_nt2 CTgene
1 TGCACGCAAGCTACACAATCCCCGTACACTTTT IGHV3-33.IGHD3-3.IGHJ6.IGHM_IGKV2-24.IGKJ2.IGKC
2 TGCACGCAAGCTACACAATTTCCGTACACTTTT IGHV3-7.NA.IGHJ4.IGHM_IGKV2-24.IGKJ2.IGKC
3 TGCACGCAAGCTACACAATTTCCGTACACTTTT IGHV3-21.IGHD3-10.IGHJ4.IGHM_IGKV2-24.IGKJ2.IGKC
4 TGCACGCAAGCTACACAATTTCCGTGGACGTTC IGHV3-53.NA.IGHJ4.IGHM_IGKV2-24.IGKJ1.IGKC
5 TGCACGCAAGCTACACAATTTCCTCACACTTTC IGHV4-34.NA.IGHJ6.IGHM_IGKV2-24.IGKJ4.IGKC
6 TGCACGCAAGCTACACAATTTCCTCACTTC NA_IGKV2-24.IGKJ5.IGKC
CTnt
1 TGTGCGAGAGATCGTCAAGTATTACGATTTTTGGAGTGGTTATTACCCTACTACTACTACGGTATGGACGTCTGG_TGCACGCAAGCTACACAATCCCCGTACACTTTT
2 TGTGCGAGAGATGGCGGTGACGACGGGTACTGG_TGCACGCAAGCTACACAATTTCCGTACACTTTT
3 TGTGCGAGAATACCTCCGGTTACTATGGTTCGGGGAGTTATTAAATTCTTTGACTACTGG_TGCACGCAAGCTACACAATTTCCGTACACTTTT
4 TGTGCGAGAGATTCGTATAGCAGCCATTACTGG_TGCACGCAAGCTACACAATTTCCGTGGACGTTC
5 TGTGCGAGTGGAATTAAGCAGCAGCTGGCCTACTACTACTACGGTATGGACGTCTGG_TGCACGCAAGCTACACAATTTCCTCACACTTTC
6 NA_TGCACGCAAGCTACACAATTTCCTCACTTC
CTaa CTstrict
1 CARDRQVLRFLEWLLPYYYYGMDVW_CTQATQSPYTF IGH:Cluster.32.IGHV3-33_IGLC:Cluster.14.IGKV2-24
2 CARDGGDDGYW_CTQATQFPYTF IGH.403.IGHV3-7_IGLC:Cluster.14.IGKV2-24
3 CARIPPVTMVRGVIKFFDYW_CTQATQFPYTF IGH.33.IGHV3-21_IGLC:Cluster.14.IGKV2-24
4 CARDSYSSHYW_CTQATQFPWTF IGH.1202.IGHV3-53_IGLC:Cluster.14.IGKV2-24
5 CASGIKQQLAYYYYGMDVW_CTQATQFPHTF IGH:Cluster.56.IGHV4-34_IGLC:Cluster.14.IGKV2-24
6 NA_CTQATQFPHF NA.NA_IGLC:Cluster.14.IGKV2-24
> #igraph
> igraph.object <- clonalCluster(combined,
+ chain = "IGH",
+ sequence = "aa",
+ threshold = 0.85,
+ samples = NULL,
+ exportGraph = TRUE)
错误于clonalCluster(combined, chain = "IGH", sequence = "aa", threshold = 0.85, :
No clusters detected with current parameters.` |
Matrix products: default locale: time zone: Asia/Shanghai attached base packages: other attached packages: loaded via a namespace (and not attached): |
|
I found that setting the exportGraph parameter of the clonalCluster function to F allows the function to run and generate the igraph.object, but it does not produce the clusters. To generate an igraph similar to the one which Nick did in https://www.bioconductor.org/packages/devel/bioc/vignettes/scRepertoire/inst/doc/vignette.html#1_Introduction, I tried using the cluster generated by combineBCR function to draw the clone connection plot, which produces a similar effect, though not as visually appealing as Nick's example. #igraph BCR connection plot
library(scRepertoire)
library(igraph)
library(ggraph)
library(tidyverse)
# Load data
csv1 <- read.csv("E:/R/filtered_contig_annotations.csv", stringsAsFactors = F)
combined <- combineBCR(csv1, samples = c("sp_exp"), ID = NULL,
threshold = 0.95)
# Create clonal graph object
igraph.object <- clonalCluster(combined,
chain = "IGH",
sequence = "aa",
threshold = 0.95,
samples = NULL,
exportGraph = F)
# Convert graph object to a dataframe
df <- as.data.frame(igraph.object)
if (!is.data.frame(df)) {
df <- as.data.frame(df)
}
# Extract IGH:Cluster part from sp_exp.CTstrict column to define each node's cluster
df <- df %>%
dplyr::mutate(IGH_cluster = sapply(strsplit(as.character(sp_exp.CTstrict), "_"), function(x) x[1]))
# Filter rows containing IGH cluster information and remove NA values
df_igh <- df %>% dplyr::filter(!is.na(IGH_cluster))
# Keep only clusters with at least 3 nodes
df_filtered <- df_igh %>%
group_by(IGH_cluster) %>%
dplyr::filter(n() >= 3) %>% # Keep only clusters with at least 3 nodes
ungroup()
# Create an edge list, connecting samples within the same cluster
edge_list <- df_filtered %>%
group_by(IGH_cluster) %>%
summarise(pairs = list(combn(sp_exp.barcode, 2, simplify = FALSE))) %>%
unnest_longer(pairs) %>%
mutate(from = sapply(pairs, `[`, 1),
to = sapply(pairs, `[`, 2)) %>%
dplyr::select(from, to)
# Remove possible NA values
edge_list <- edge_list %>% dplyr::filter(!is.na(from) & !is.na(to))
# Create graph object, only keep nodes that are in the filtered edge list
graph <- graph_from_data_frame(edge_list, directed = TRUE)
# Update graph node attributes
V(graph)$cluster <- df_filtered$IGH_cluster[match(V(graph)$name, df_filtered$sp_exp.barcode)]
V(graph)$size <- 1.5 # Node size
# Assign a unique color to each cluster
library(viridis)
unique_clusters <- unique(V(graph)$cluster)
color_map <- setNames(viridis(length(unique_clusters)), unique_clusters)
col_samples <- color_map[V(graph)$cluster]
# Use MDS layout for graph plotting
layout <- layout_with_mds(graph)
plot(
graph,
layout = layout,
vertex.size = V(graph)$size,
vertex.label = NA,
edge.arrow.size = 0.1,
vertex.color = col_samples,
vertex.frame.width = 0.2
) |
Per issue #433 exportGraph = TRUE breaks in the context of a single sample, using igraph::union before evaluation statement
|
Thank you so much for the excellent followup and rundown of the problem. I was able to find the issue - evaluation statement during exportGraph = TRUE assumed that there would be more than 1 sample - so feeding in a single sequencing run broke it. We now combine the entire graph before we evaluate: scRepertoire/R/clonalCluster.R Line 134 in f9a4a12 Here is a rundown of the fixed function: 
Fix is up in the dev branch - @Qile0317 and I are working on some minor changes before we get this version shipped and pushed to bioconductor/main branch of github Thanks again, |
|
@ncborcherding Haiqiao |
My combined <- combineBCR(csv1, samples = c("sp_exp"), ID = c(NULL)) successfully created Cluster in the CTstrict column with the default threshold of 0.85, which seems perfectly correct.
combined[["sp_exp"]][["CTstrict"]]
[1] "IGH.1376.IGHV3-43_IGLC:Cluster.12.IGKV2-24" "IGH.1376.IGHV3-43_IGLC:Cluster.12.IGKV2-24"
[3] "IGH.1505.IGHV1-18_IGLC:Cluster.12.IGKV2-24" "IGH.1741.IGHV3-64D_IGLC:Cluster.12.IGKV2-24"
However, when using the clonalCluster function as follows:
igraph.object <- clonalCluster(combined,
chain = "IGH",
sequence = "aa",
threshold = 0.85,
samples = NULL,
exportGraph = TRUE)
it returns an error: clonalCluster(combined, chain = "IGH", sequence = "aa", threshold = 0.85, : No clusters detected with current parameters.
Do you have any ideas about this? Thank you.
The text was updated successfully, but these errors were encountered: