Here is a small survey. I'm wondering if anyone here is performing data analysis for immune repertoire sequencing data (or should we call it immune metagenome sequencing :). I know that it a developing field and there are few labs that are involved. Just wanted to estimate the real extent of this kind of bioinformatics research. So let me know if you or someone your know are involved in this kind of research. Please also tell what kind of sequencing platform, library preparation technique and analysis software (or in-house scripts) are you using.
P.S. The term immune repertoire sequencing here refers to sequencing of antigen receptors of T- and B-cells.
Thanks in advance
Update
A dedicated immune repertoire sequencing community forum can be accessed at http://b-t.cr/
Well technically those are different things. HLA typing is about variant calling and phasing, while immune repertoire sequencing is about extracting and assembling an enormous number of highly variable (CDR3) sequences of immune receptors. Ideally those two approaches should complement each other, as certain immune receptor variants have different functions depending on HLA context, e.g. several autoimmune diseases are known to be linked to specific HLAs, as well as specific Variable segments of immune receptors
Are they going to do 16S profiling to achieve it? This sounds interesting, still I realized that I've not quite clearly described the topic itself. Microbiome profiling with 16S is technically the same task as T-/B-cell repertoire sequencing, both use sequences of hypervariable regions to distinguish unique OTUs / clonotypes (T-/B-cell families sharing the same antigen receptor). While OTUs are more or less well-annotated, the T-/B-cell clonotypes are not. Only a tiny fraction is known to be associated with response to certain pathogens, tumor-associated antigens and autoimmune diseases. There is also no comprehensive curated databases with annotation of immune clonotypes to the best of my knowledge.
We are interested in sequencing and studying rhesus macaque B cell repertoire. We have used 454 platform before but now moving to RACE PCR for 'unbiased' Ig gene amplification and miseq 2X300bp platform for sequencing. For analysis we so far have used usual tool/servers such as vquest, igblast, vdjfasta, igAT, Abmining toolbox etc. I see you have developed set of tools for Ab gene analysis, we would be interested to try your vdj tools on our data. One of your tools, IgWrapper seems can analyze only human, mouse seqs at the moment, I wonder if its possible to make a rhesus Ig database and incorporate in igwrapper in order to analyze rhesus vdj seqs ? I also agree that igblast miss few analysis parts like CDR3 info etc and outputs are not user-friendly compared to imgt-highvquest. Igblast currently using our rhesus germline database for vdj annotation of user-submitted sequences. I think it will be interesting to use igblastwrapper because it includes additional analysis features, would you be interested to help us adding the rhesus germ db in your tools? Thanks!
ADD REPLY
• link
updated 2.8 years ago by
Ram
44k
•
written 10.1 years ago by
GP
▴
10
0
Entering edit mode
Hello and thanks for the feedback!
Basically the IgBlastWrapper was developed as a quick solution for the incoming 300+300 data. It relies on the data provided with original IgBlast (except for position of conservative "reference" codon in the Joining segment). Actually the IgBlast team released monkey ig support this july, which I've totally missed (IgBlastWrp was not updated since the end of June). I'll now start to integrate monkey sequences, it will take me some time to make Joining segment references and test all the stuff, I'll update when I've finished. Note that most likely I'll request a sample of 1000-10000 reads for testing purposes, as I'm limited to human IG seq here.
ADD REPLY
• link
updated 2.8 years ago by
Ram
44k
•
written 10.1 years ago by
GP
▴
10
0
Entering edit mode
Update! Here is the new release with additional species support https://github.com/mikessh/igblastwrp/releases/tag/v0.4. Please test it with your rhesus sequences, as I only tested it using human data. Also note that it is quite slow, while given MiSeq provides around 20mln reads it is not very problematic, especially if the oversequencing is complemented with using UMI (unique molecular identifier tags), which shrinks the data to 2mln consensuses.
Thanks very much, I'll try it later (we currently in the process of producing our 1st miseq dataset, we may try on our already published 454 data though, will update you!
ADD REPLY
• link
updated 2.8 years ago by
Ram
44k
•
written 10.1 years ago by
GP
▴
10
This is Peter Che, from a Japan Repertoire Genesis, Inc. We are dedicated to providing the service for immune repertoire analysis. Please visit our HP (http://www.repertoire.co.jp/en/index.html). And please don't hesitate to contact me if you need any further information or my assistance.
Thank you for taking time to read this.
With warm regards,
Peter Che
ADD COMMENT
• link
updated 2.7 years ago by
Ram
44k
•
written 8.0 years ago by
che
▴
20
We are interested in sequencing and studying rhesus macaque T cell repertoire. We have used Illumina platform and prepared library from PCR amplicons of rhesus macaques genomic DNA between TRB V and J. We don't analyse data ourselves but collaborate with other bioinformaticians, so we don't know what software they used. I see you have developed MiTCR and MiGEC, we would try them on our data (but MiTCR don't support monkey sequences). We would like you can add macaque data into MiTCR very much.
There is a lab in my institutes which are study BCR of monkeys. they use iCubate platforms and do some in-house scripts to analyse their data.
Just replied you regarding monkey TRBVJ segment references here. It seems that in RepSeq many more people work on organisms other than human and mouse than I would expect.
Could you tell me more on the iCubate platform? From my first impression this is an qualitative PCR on some common pathogens. But I think this would be very powerful for the annotation of samples that could then integrated in BCR analysis.
Thank you for your replies in this post and Mitcr: A Software Tool For Analyzing T-Cell Receptor Sequencing Data. Yes, there are many people work on organisms other than human and mouse. Especially rhesus monkeys, which are an important animal model for many human diseases, most time best and only one. So you should include this species into your future developing software.
iCubate perform PCR in a close cassette to prevent contamination. The company provide a pool of primers amplifying TCR or BCR and permorm multiplex PCR. The products can be sequnced directly on NGS. The company said their primers introduce the least bias in PCR.
ADD REPLY
• link
updated 2.7 years ago by
Ram
44k
•
written 9.9 years ago by
mcchance
▴
30
0
Entering edit mode
Hi!
We are working on reference issue, however it turns out to be a complex one. IMGT references were manually built using GenBank, and mapping them back to genome assembly reveals lots of issues with them. We're now trying to move on to purely genome-assembly-based V/J segment coordinates, which are far more robust. Frankly speaking, the situation with references in our field would be ok for 1990th, but not for 2014.
I also believe that the pool of V/J alleles should be re-checked using full-length high-throughput TCR/IG sequencing.
iCubate: I see, actually we are mostly using RACE-based technique, which naturally removes all the biases, and we use paired-end sample barcoding to prevent contamination due to PCR chimeras. I just wonder what is the length of the amplified PCR product? Were you able to get full-length BCR with them?
Thanks for opening this discussion. I am new to Rep-Seq but interested in getting start to know all these. One question though after conducting some research:
I saw lots of people using single cell sequencing (Smart-Seq) to sequencing T / B cell. I assume this is for evaluating the abundance of Immune repertoire of a organism and sequencing single cell would be the most precise way to achieve this. So am I right that most Rep-Seq is about sequencing single cell and conducting analysis on TCR gene for abundance. Right?
I am not familiar with immune repertoire, but am working on HLA a bit. Not sure if/how the two are related.
Well technically those are different things. HLA typing is about variant calling and phasing, while immune repertoire sequencing is about extracting and assembling an enormous number of highly variable (CDR3) sequences of immune receptors. Ideally those two approaches should complement each other, as certain immune receptor variants have different functions depending on HLA context, e.g. several autoimmune diseases are known to be linked to specific HLAs, as well as specific Variable segments of immune receptors
Thanks for the background. This stuff sounds familiar from my Immunology class a million years ago, and sounds incredibly complex.
I am at a summer school where I have learned that the CDC will mandate NGS type sequencing as a replacement for detecting pathogenic contaminations.
Are they going to do 16S profiling to achieve it? This sounds interesting, still I realized that I've not quite clearly described the topic itself. Microbiome profiling with 16S is technically the same task as T-/B-cell repertoire sequencing, both use sequences of hypervariable regions to distinguish unique OTUs / clonotypes (T-/B-cell families sharing the same antigen receptor). While OTUs are more or less well-annotated, the T-/B-cell clonotypes are not. Only a tiny fraction is known to be associated with response to certain pathogens, tumor-associated antigens and autoimmune diseases. There is also no comprehensive curated databases with annotation of immune clonotypes to the best of my knowledge.
Hi Mikhail,
We are interested in sequencing and studying rhesus macaque B cell repertoire. We have used 454 platform before but now moving to RACE PCR for 'unbiased' Ig gene amplification and miseq 2X300bp platform for sequencing. For analysis we so far have used usual tool/servers such as vquest, igblast, vdjfasta, igAT, Abmining toolbox etc. I see you have developed set of tools for Ab gene analysis, we would be interested to try your vdj tools on our data. One of your tools, IgWrapper seems can analyze only human, mouse seqs at the moment, I wonder if its possible to make a rhesus Ig database and incorporate in igwrapper in order to analyze rhesus vdj seqs ? I also agree that igblast miss few analysis parts like CDR3 info etc and outputs are not user-friendly compared to imgt-highvquest. Igblast currently using our rhesus germline database for vdj annotation of user-submitted sequences. I think it will be interesting to use igblastwrapper because it includes additional analysis features, would you be interested to help us adding the rhesus germ db in your tools? Thanks!
Hello and thanks for the feedback!
Basically the IgBlastWrapper was developed as a quick solution for the incoming 300+300 data. It relies on the data provided with original IgBlast (except for position of conservative "reference" codon in the Joining segment). Actually the IgBlast team released monkey ig support this july, which I've totally missed (IgBlastWrp was not updated since the end of June). I'll now start to integrate monkey sequences, it will take me some time to make Joining segment references and test all the stuff, I'll update when I've finished. Note that most likely I'll request a sample of 1000-10000 reads for testing purposes, as I'm limited to human IG seq here.
Ok, Thank you for your reply.
Sure, we can provide you macaque sequences.
Update! Here is the new release with additional species support https://github.com/mikessh/igblastwrp/releases/tag/v0.4. Please test it with your rhesus sequences, as I only tested it using human data. Also note that it is quite slow, while given MiSeq provides around 20mln reads it is not very problematic, especially if the oversequencing is complemented with using UMI (unique molecular identifier tags), which shrinks the data to 2mln consensuses.
Thanks very much, I'll try it later (we currently in the process of producing our 1st miseq dataset, we may try on our already published 454 data though, will update you!