jcao89757 / tessa Goto Github PK
View Code? Open in Web Editor NEWMapping the Functional Landscape of TCR Repertoire
Mapping the Functional Landscape of TCR Repertoire
The gene expression value was scaled,log-transformed? How to obtain the suitable expression value for TESSA from UMI count?
Hi, Dear TESSA developers,it is a good idea for clusering single cell by joint TCR and RNA profile,,but Tessa takes too long to run,My guess is that this is caused by 1000 iterations of MCMC,So I would like to ask how the number of iterations affects the final result. Can I customize it?How much do you recommend?Thank you and I am looking forward to your reply.
Greetings again! I have successfully run this analysis on ~30k T-cells without any problem. Now, I am attempting to run this analysis on a dataset of ~100k T-cells and I am receiving the following error:
Error in hclust(dist(t(t0))) : size cannot be NA nor exceed 65536
Calls: Tessa ... initialize -> initialize_cluster -> cutree -> nrow -> hclust
Execution halted
I did some testing and if I subset my data down to 100 cells this error does not arise. I then looked into the hclust function and it seems that it will not accept a distance matrix that has more than 65536 observations? Is my interpretation of this error correct? If it is correct, then I suppose that I will need to implement a different clustering function during this step?
Thanks for your help!
Hi, I'm fresh in sc-TCR related analysis, glad to see this new tool.
As I know, to decide a clonotype of one T cell, one ’TCRα + TCRβ pair‘ profile should be necessarily provided. But some sc-TCR analysis based on 10x genomics vdj technology which has much lower rate to detect TCRα have to use TCRβ only. I'm not sure about how many false-positives it may bring in. e.g. x50 copies of one TCRβ type are detected in two different tissues, could say these two tissues both have this kind of T cell, but if add TCRα result, might find that no 'α+β pair' are overlapped between them.
My question is that: have you thought about this issue, and would 'using TCRβ only' affect tessa method like correlation of TCR type x Tcell expression or sth. else ?
what does node size represent in the figure generated from plot_Tessa_clusters? How to understand the figure?
Hi Ze,
Different lengths of CDR3 sequence can be solved by Tessa, a wonderful tool. I have an Error when I used my data running TCR network construction. The error is as followd:
Initialization
[1] 123
Error in t[, !duplicated(cdr3), drop = F] :
(subscript) logical subscript too long
Calls: Tessa -> initialize
Execution halted
embeding.csv have accomplished, but I cannot find out where is the error from. Could you give me some advice? I can run to completion the example, but the version is not completely the same as you. I dont know whether this mistake is caused by version.
Thanks a lot.
Ren
Hi @jcao89757 , thanks for the great package and all your efforts in maintaining it!
I ran the relined code: “python3 Tessa_main.py -tcr ./example_data/example_TCRmeta.csv -model ./BriseisEncoder/TrainedEncoder.h5 -embeding_vectors ./BriseisEncoder/Atchley_factors.csv -output_TCR test.csv -output_VJ testVJ.csv -output_log test.log -exp ./example_data/example_exp.csv -output_tessa /path/to/tessa_results/ -within_sample_networks FALSE”
and the error was as follows: “ERROR: please input valid TCRs/embedding.”
This code generates the embedded TCRs file (test.csv), but the running results are inconsistent with your example results, as follows:
Thanks,
Ze Zhang
Dear all,
I tried the Tessa tool and found in the post-analysis step, library ggplot2 is missed in function plot_tessaClsuters. This gives an error for R plotting. What's more, the error message 'ERROR: please input valid TCRs/embedding.' is confusing.
Best!
Hi Ze,
The code can change the CDR3 sequences to a 30-d numerical vector. I want to know how to find each or some numerical vector corresponding amino acid sequence like Extended Data Fig.1 a decoder.
Thanks a lot
Ren
The printout of TESSA is "cluster rate" and "b acceptance rate", but if I would like to get the cluster purity that you've mentioned in your paper, how can I get it? Thanks.
I want to train my data using the autoencoder ,In the script "encodeTCR_CNN.py", in the step: convert to 3D array it called:
cannot reshape array of size 11109250 into shape (120,5,1),
Could you please tell me why this error occurred?
Greetings, brilliant tool and really excited to implement this. I was wondering if there was any documentation on the Rdata output? Specifically, I am assuming that the tessa_results$dt contains the TCR distances for each network, but I am unclear on the structure of it.
For example, I would like to calculate the unexplained variance for some of my networks, which I think I can derive from this Rdata output, but I am just not sure how to get there without any explanation of the output structure.
Thanks so much for your time!
what does node size represent in the figure generated from plot_Tessa_clusters?
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.