Scientist Stories: Long Xu, Platforms for Antibody Discovery

A good morning or afternoon to you. Thank you for joining us. I'm Jan Breker, Chief Operating Officer of the Antibody Society. Today's webinar is one in a series designed to inform and educate our members as well as the broader scientific community about topics relating to antibody discovery and development. Our expert speaker, Dr. Long Xu, will discuss the pros and cons of innovative technologies and platforms that can be applied to antibody Discovery. Please note this webinar is being recorded. Please do add any and all questions to the Q&A box in the viewer, and those questions will be answered at the end of the webinar. Without further ado, I'll now turn the show over to our speaker.

Hi everyone, and I would like to thank the Antibody Society for the opportunity to speak here., and a special thanks to Janice and her colleagues for their tremendous help. And today I'm going to share about new techs and platforms to empower antibody discovery. Over the past 3 decades, the development of antibody drugs are rapid and promising, and we are seeing more and more antibody drugs entering the market. And in 2021, FDA approved the 100th antibody drug. And now we are entered the era of 100+ antibody drugs. And antibody drugs are gaining more and more share of global pharmaceutical market. And this table lists top 10 best-selling drugs in the first half of 2023, you can see that out of top 10, 7 are antibody drugs, including all the top 3. And that's phenomenal. And why antibody drugs or biological drugs have such good performance? The reason is that biological drugs You know, they have some unique and special advantages over small molecule drugs, such as the beta affinity, beta binding specificity, and the beta tissue selectivity, and so on. And all these are increasingly favored in clinic applications. Okay, the trend looks very good and encouraging, but it is far from perfect. And we all know that the antibody industry has been constantly troubled by a series of bottlenecks such as big size, poor tissue permeability, and difficulty in modular applications. And, you know, many antibodies have low affinity and not high enough to become drugs. And also we came across a lot of problems associated with immunogenicity. And these bottlenecks are calling for new options to break through. And we think that single domain antibodies may be a new option. And I assume, you know, most of our attendees have a rough idea about the chemistry that you write single domain antibodies. You can see the picture shows the comparison between conventional antibody and and camelid-derived heavy chain only antibody. Okay, so it's interesting that we found that in camelids, you know, in their blood, there are naturally a lot of heavy chain only antibodies with light chains missing. And the variable domains of these heavy chain antibodies are called VHHs or narrow bodies because their size are just in nanometer range, and which offers them a lot of unique features such as good stability and good permeability. Okay, more than that, say you can see that they are pretty stable in some very extreme conditions And they have, you know, great potential, you know, for as a modules for downstream engineering. And they have very good solubility, and they have higher homology to human antibody heavy chains. And also they are compatible with a variety of production systems. Okay, and these advantages, you know, enable the nanobodies in a wide range of applications, such as therapeutic antibodies or diagnostic tools. Say, the nanobody can be used to construct bispecific or multispecific antibodies. They can also form the conjugates with toxin payloads or enzyme or radioisotopes, and they can be used in CAR-T or nanoparticle drug delivery systems. Okay, at the same time, they, they have great potentials to use in molecular diagnostics or as imaging reagents, you know, due to their good properties such as deep tissue penetrations and a quick clearance from blood. Okay, since VHH or nanobody is so good, and how to use it? Here you can see a workflow. And first of all, we have to do antibody discovery. The starting point is to immunize commonly the species and generate the antibody library. Then screen hits from the library against a given antigen. Then perform function assays on the hits to identify lead antibodies with application potential. Next is do the lead optimization including humanization and affinity maturation. Okay, after these steps, and hopefully the late antibody, you know, meet all the requirements, and so they can enter the final stage application. And actually, there are many methods for animal immunizations and antibody screening. And today I will mainly talk about phage display technology. Okay, this picture will give you a brief demonstration of phage display used in narrow-body discovery. Phage display is a reliable method for narrow-body discovery, and And actually, we begin with, you know, the alpaca immunization with a given antigen, and then we obtain PBMC from the alpacas, and then enriched B cells capture its messenger RNAs, reverse transcribe it into cDNA library, and then the VHH genes got PCR amplified and cloned into phage meats. And then the phage meats transformed into competent cells so that, you know, the phage can display the VHH on their surfaces. And this method allows for the construction of phage libraries capable of expressing a vast connection of VHH variants. The library size is typically more than 10 to the 9th power. But actually, what we want is just dozens of hits. Okay, that's— you can see the huge difference from 10 to the power of 9 to the dozens of hits. It's clearly very difficult. It's like, you know, finding a needle in a haystack. Okay, but, you know, we just use a very simple antigen-antibody binding mechanism. Okay, so we run several rounds of panning to isolate, you know, the good hits. Okay, so what, what, what do we do? Okay, so in each round, phages that can bind specifically to the given antigen are isolated and amplified, while those phages don't bind antigens get washed out. Okay, so in this way the positive phages that has desired binding properties are dramatically enriched in the library. Okay, so usually after 3 rounds of panning, the positive rate can reach double-digit percentages. At this point, we can pick monoclonal phages from the library to do the ELISA binding assays to identify the positive phages and then sequence them to obtain the VHH gene. Okay, next I will share with you a complete case from the narrowbody, you know, discovery until the applications. And in this case, The antigen is called X. It's actually a cell surface receptor expressed on T cells, and it is also a co-stimulator for T cell activations. Okay, so this just shows the immunization is pretty good. You know, the alpaca got well immunized, and the title looks very good after the second round of immunization. So we collected the PBMC after the second round of immunization for the construction of the phage library. And here are some QC steps, and you can see some tests done during the construction of the VHH library. Such as insertion rate, library size, and diversity of the VHH sequences. You can see that our library size is 10 to the power of 9, and the diversity check shows that, okay, not many duplicated sequences. So that means the diversity of our phage library is pretty good. Okay, so in the panning stages, we use both solid phase and liquid phase panning. Okay, this is a solid phase panning. You can see that the titer of the phage pool improved greatly after each round of panning, and that indicates that the enrichment of the positive phage is great. And then we selected— we randomly selected about 1,000 clones, and finally about 360 clones are identified as positive clones. And then after sequencing, 11 unique sequences are obtained. This shows the results from the liquid phase panning. It's similar. You can see that, so out of thousands of clones selected, and 952 clones are identified as positive, and then we got 43 unique sequences. So totally From both liquid and solid panning, we got 54 unique sequences. So that's our 54 hits. Then we performed a series of functional assays, you know, to characterize those 54 hits, you know, to identify the good late antibodies with good application potentials. The first assay is ELISA binding assay. Okay, it's pretty simple, just, you know, to check if, you know, these kits can specifically bind to the antigen X. Okay, the pictures, the results are here. You can see you know, the binding curves are pretty good actually. Out of 54 hits, 35 antibodies show very good binding curves. And then all these hits are subjected to very accurate affinity data measurement on Biarcou platform. Here just to show some BL-Co graphs, okay. And actually 49 out of 54 hits show clear, you know, binding against the eggs in SPR test. And the affinity of the hits ranges from 10^-7 10^-10. Okay, you can see the bar graph here. Actually, you can see that most of our hits, their affinity falls into the range of 10^-8 power or better. So that is pretty good. Okay, so here we're trying to examine, you know, the binding properties of our hits against the cells. Okay, in the previous slides you see they have good binding properties against the antigen proteins. Now we want to examine, you know, the binding properties on the cell level. Okay, so all these kids are examined by the flow cytometry against CHO-K1 cells overexpressing the antigen X. And the binding curves are shown here. You can see that Out of 54 hits, 43 VHH antibodies bind against the CHO-K1 cells overexpressing antigen X. So that means, you know, most of our hits, they have very good binding properties against the antigen X, even, you know, at the cell level. That's an interesting assay. It's used to examine the antibody's agonist activity. Okay, actually here we use a luciferase reporter system. Okay, so the mechanism behind it is, you can see the picture actually. So when the the, the, when the antibody binds specifically against the antigen X, you know, on the cell surface, that can, you know, that signal can promote the production of luciferase, and which reacts with its substrates to produce luminescence. Okay, and the intensity of luminescence is positively correlated with the antibody's agonist activity. Okay, so that's interesting. And let me see the performance of our hits and have a look at this bar graph. The blue bar are the positive control you can see the lumination intensity is high. The red bar are the isotope control, okay? And all these green bars are part of our hits, okay? And we, so when the lumination intensity is greater than 1.5 times of isotope, we can say that, you know, the agonist activity is clearly over there. So we checked all our hits, and actually there are 31 out of 53 hits show clear agonist activities. Okay, now then, this assay is used to determine the ability of the VHH to block interaction between the ligands and the cell membrane-bound antigen X. So the results are still satisfactory, and out of 51 hits used in these assays, actually 44 antibodies show blocking effect. That's a big part. Okay, so after screening and characterization of all these hits, and we actually selected several leads and performed the further optimization to meet the objects, the objectives of the project, and then they are ready for the applications. In this case, they are used in the construction of therapeutic bispecific antibodies, and actually we are able to construct a variety of common forms of by specific antibodies. Okay, so the above case is pretty straightforward, and you can see the results are satisfactory and everything seems very smooth. But actually, in our daily work, we come across a lot of challenges. Here's just one, say, what if it is difficult to prepare soluble proteins and antigens? You know, actually in the previous case we use soluble protein eggs as antigens, so it's pretty straightforward and simple. But in many cases, okay, such as transmembrane targets or GPCR targets, we cannot prepare soluble proteins and antigens because once this antigen is detached from the cell membrane, it loses its proper structures and functions. Okay, that's a huge challenge. What should we do in this case? Okay, here is another case, and the goal of this project is trying to identify the Claudin 18.2 antibodies. Okay, oh, we know that Claudin 18.2 cannot be prepared as soluble protein, so in the animal immunization In the first 2 runs of immunization, we use a DNA plasmid encoding Clodin-18.2 as antigen. And actually, you can see the titer of the first 2 rounds, not that good. Then in the later 3 rounds of immunization, we use CHO-K1 overexpressing cell line as antigens. Okay, and the title of the— in the title you can see a big jump, you know, especially after the 5th round of immunization. Okay, so at this point the immunization is good enough for the construction of phage library. And then we performed the 3 rounds of panning. After each round of panning, we randomly selected 88 clones for the expression and the validation. And so the case is pretty lucky actually. Even after the first round of panning, we obtained some positive clones and unique sequences. So totally, from 200-something clones, we got dozens of positive clones and actually 13 unique sequences. And we expressed these hits for validation. Finally, we selected the top 4 hits. You can see the lower left, the graph just shows the top 4 hits and their binding curve against the the cells, the cells overexpressing protein 18.2. Okay, so clearly the binding properties is what we desired. And at the same time, we run some control experiment and say when the blank 293 cells used and there's no binding at all,. And the lower right picture, you can see that here's the 293 cells overexpressing Cladin18.1 is used, and there's no binding at all. That's pretty good because we know that actually between Cladin18.1 and Cladin18.2, actually there's only 7 amino acid differences in ECD. But, you know, on our platform, we successfully, you know, identified and isolated such good hits that can specifically bind against Claudin 18.2. Okay, here's another challenge. People may ask, so can we use a naive phage display library instead of the alpaca immune antibody library? The answer is a big yes because we have our own alpaca range. We have more than 300 alpacas. So based on 300 alpaca donors, we collected their PBMCs and using phage display technology and constructed the huge naive alpaca phage display library. And our library size is more than 10 to the power of 12. That's pretty huge. And here is a case, you know, we just use our naive alpaca antibody library directly for screening, for antibody screening. Okay, so actually we performed 3 rounds of liquid phase panning. After each round of panning, we randomly select 88 molecular clones. And finally, so out of 264 clones, exactly 100 positive clones are identified. And these 100. These 100 positive clones are sent for sequencing, and finally we got 18 unique sequences. And we expressed all these 18 hits, and here are the top 5 hits. And you can see their affinity is pretty good, some are even better than the positive control, the the red curves. Okay, so that means, you know, with naive library we can successfully screen for, you know, the nanobody with desired binding properties. And this slide just shows some statistics results. Okay, we collected the 600-something unique sequences we deliver to our clients. And it is interesting that actually one-third of our hits, their affinity are in the 10^-8 power range. Another one-third is just in the nanomolar range, and some are even better than that. So totally, we can say that 80% of our BHH hits have the affinity of 10 to the minus 8 power or even better. So that means our library, our immunization, and our phage display display technology are pretty good to screen for quality VHH hits. Okay, then I would like to say something about, you know, the antibody optimization, because we know that usually the VHH or nanobodies They're pretty good, but usually their affinity are not that high enough to become drugs. Usually their affinity is in 10 to the minus 8 power. That's not good enough. So next I will introduce our proprietary affinity maturation platform. We call it MC-MS., and it's full coverage mammalian expression system for affinity maturation. Our platform features full coverage and no bias. And how we do that? This workflow gives you a rough idea. We know that, you know, as to the antibodies' binding properties, it's mostly determined by its CDR regions. So we focused on the CDR regions, and for each amino acid position in CDR regions, we generate a site-saturated mutant plasmid library. The library covers 17 single-point mutations at this position. Y17, because, you know, the original amino acid methionine and cysteine, these three are excluded. Okay, so then the mutant plasmid library is transformed into competent cells, and the cell culture is then sprayed on agar plates. Each agar plate corresponds to one specific position in CDR regions. Okay, and the clones on each agar plates covers all 17 mutant plasmids at that position. Oh, then we randomly select 88 clones from each agar plate and cultured in 96-well plates for plasmid amplification and extraction. Why 8x? Because based on our experience, 8x clones may cover all 17 mutant plasmids, which is confirmed by the sequencing results. You know, on the right you can see that that's just the sequence results from one plate. You see the 8x plasmid cover all 17 mutant plasmids. And then the plasmids are moved from the 96-well plate to another 96-deep well plate to express the antibody in our high-throughput expression system. Okay, I want to, uh, to point out that so up to this point, for each amino acid in CDR region, we end up with 196 plates, with each well containing antibody supernatant expressed by a single point mutant plasmid at that position. So you can see the workload is pretty huge because, say, if there's 70 amino acid positions in CDR regions, we end up with 70 plates. That is over 6,000 wells of antibody supernatants. Okay, so with so many wells of supernatants, we can do the high-throughput ELISA screening assays, okay, to identify hotspots. So what are hotspots? Hotspots are reference to the mutations to the wells, you know, well, the supernatants show significantly higher affinity than the parental antibody. Okay, so once we identify the hotspots, we can go back to find the, you know, the corresponding plasmid and do the sequencing to identify exactly what mutation at that position caused such affinity increase. So after the ELISA assays, we collect abundant hotspot information. Then we select couples of hotspots and then do the combination and design dozens of candidate antibody sequences. These dozens of candidate antibody sequences I got will be expressed and purified and then subjected to SPR test on biaco to make sure that, you know, we finally, we got the antibody sequences that have, you know, significantly increased affinity. Okay, I hope you can understand, you know, the story behind this workflow is a little complicated. Okay, then I just, you know, go through some cases of affinity maturation. As I just mentioned that nanobodies usually have poor affinity. Say here are two examples, you know, they are all in the 10���� range, and the binding rate on the left is pretty slow, while the dissociation rate on the right is too fast, and what should we do? Okay, so for the nanobodies with poor binding rate, we need to improve its K off. And on our platform, we identify these 5 hotspots and we do the combinatorial expression. And here are top, you know, candidate antibody sequences and show very good results. We can see from the lower graphs that after the affinity maturation, the binding curve of the antibodies are much better than the parental antibodies. And we also see the data on the upper right corner. And yeah, actually, after the optimization, the antibody affinity increased up to more than 8-fold, and we can clearly see that, you know, their K-on increased, you know, very significantly. Over here is the second case that we, uh, very— the dissociation rate is too fast, then we have to lower down its K-off. Okay, so in this case, on our platform, we identify These top 4 hotspots, we do the combination and design the candidate sequences and do the expression. So here are 3 sequences with good performance. We can see that, you know, from the graph that the dissociation curve improves a lot and the affinity of the antibodies increased from 6-fold to 23-fold. And we can also clearly see that the value, the KD value, decreased greatly after the optimization. And here's another case that's a big project, you know, starting from the narrow-body discovery, and then followed by humanization. Okay, and then after our affinity maturation, we can see that the improvement of the affinity is tremendous and is up to, you know, more than 260-fold. That's a very successful case. Okay, so I've just talked about a lot of cases of antibody discovery by the phage display technology, and some of our attendees may argue that the phage display approach is, it's not bad, But it's still very, you know, time-consuming and labor-intensive. And can we have a better option? And say, can we have a, you know, a magic microscope? You know, with that we can have a close look at each B cell and tell if the antibody, you know, secreted by B cell has the desired antigen binding properties. I don't know if there's such a magic microscope in the world, but we actually have antibody high-speed sorting platform, which has very similar magic power, and I would like to share with you. And actually, this picture just to show you the sorting process. Actually, we begin with a mouse immunization, And then actually the B cells are obtained and enriched from the spleen of immunized mouse. Okay, these enriched B cells are added to the microfluidic platform for sorting. The platform can generate millions of picoliter volume droplets, very tiny, and each droplet contains up to one B cell. Okay, up to one B cell together with antigens and secondary antibodies. Okay, then each droplet will pass through the detectors. The detector can tell, you know, which droplet contains a good B cell. What is good B cell? Good B cell means that, you know, the B cell inside can secrete antibody with desired binding properties to the given antigen. Okay, well, that sounds magic. And how can the— our detector tell it is good or bad? Okay, the, the criteria for judging is simple. Actually, they are based on the specific rate of fluorescence signal. Okay, and this, the, the Radar fluorescence signals confirms that such a sandwich structure has been created in the droplet. You can see almost at the center of the picture that it shows you the sandwich structure, okay, which includes the, you know, the antibody secreted by the B cells, and the antibody binds to both antigen and secondary antibodies. Both antigen and secondary antibodies are fluorescently labeled. Okay, so interestingly, when the antigen is excited, its emitted light, fluorescent light, will excite the secondary antibody. The secondary antibody will emit that specific red fluorescent signal. Okay, okay, so it is called the FRET effect. Okay, so when a droplet with FRET effect passes through the detector, a significantly enhanced radiofluorescence signal is detected. In this way, good B cells can be collected and bad B cells go to west direction. That's a very smart design. Okay, so we have collected good B cell pool, and what should we do next? And next, on a micro well plate, each good B cell is paired with a barcoded bead. You can see that the orange one is is a barcoded bead, and the blue one, the blue one is a good B cell. Okay, the bead is covered with millions of primers to capture the messenger RNA released from the lysed B cell, which is then, you know, reverse transcribed and amplified and followed by the sequencing and bioinformatics analysis. And, you know, which results in tons of sequence information. Okay, but people may ask, okay, you get tons of, you know, sequence information of heavy chains and light chains, but how do we pair them together correctly? Okay, here is another smart design. Actually, on each barcoded beads, you know, the primer carries unique barcode. And the smarter design is that the primers on the same beads carries the same unique barcode, while the primers on different beads, you know, carries different barcode. So we know that when the, the, the, the, uh, uh, messenger RNA or the cDNA or the PCR amplification products, when, if they come from the same, the same lysed B cell, they must carry the same barcode. So with that unique barcode information, we can correctly pair cognate light chain and heavy chains. Okay, so this high-speed, you know, antibody sorting platform greatly lower down the, you know, or speed up the antibody discovery. Usually, you know, we know that the phage display platform costs us, you know, several months. But here, you know, with that single B cell sorting platform, the sorting process can be done within 1 day if everything goes smoothly. That's phenomenal. And this picture just to show typical results, you know, and you can see that we got many hundreds of sequences, you know, from the good B cells, and in each light chain and heavy chain are paired correctly based on their unique barcode information. Okay, I would like to share this case with you. It's just, you know, it was done on our high-speed single B-cell antibody sorting platform. It's used to screen for PD-1 antibody. And actually, our single B-cell sorting platform returns hundreds of sequences to us, and we selected the 62 hits, okay, and then do the antibody expression and validation. And here, just one of the assays, that's it's an ELISA binding assay. You can see that most of the hits show very good binding against PD-1, and the barcode show here actually 59 out of 62 show positive binding against PD-1. That's pretty good. And then the affinity is also very good, and most of them are in the range of 10 to the minus 9 power, that is nanomolar range, and some are even better. Okay, then we performed this ligand blocking assay, and we also find that 8 hits show very good blocking effects. Okay, so that's interesting results actually, it's done by our new platform, it's called the Katerra. Katerra is kind of the enhanced version of Biaco. And we do the high-throughput SPR-based aptamer binding. And aptamer binding can tell us, you know, can group the antibody which show similar binding epitope. And it's interesting because we know there are many PD-1 antibodies available on the market. We just chose two, Bisimilar 1 and Bisimilar 2. And we use these two together with our 60-something hits and then run the epitope binding by CUTERA. The results are interesting, and actually 19 of our hits show very similar epitope, you know, as Bisimilar 1, but none of our hits show a similar epitope as Bisimilar 2. That's interesting. And we also, this on the left, you can see the phylogenetic tree that just indicates the diversity of our hit sequences. You can see that these trees show that the diversity of our hits is pretty good. Okay, so here's That comes to the end of my presentation, and it's my great honor to have the chance to share with all the attendees. So any questions and comments are welcome. And yeah, actually, due to the time limit, you know, some details are not expanded, so you are welcome to contact us and welcome to discuss with us. I have my email address over there, and you can also browse our website.

Okay, thank you very much for that presentation. We do have a few questions, so we've got a couple of minutes here to have you answer those questions. What is the typical period of time for nanobody discovery with phage display platform at your company?

Okay, yeah, many people are interested in the timelines. Actually, I would say it's about— usually it's a typical discovery project costs less than 4 months, and Okay, the most important and time-consuming part is alpaca immunization. It usually costs, say, about 2 months because we have to, you know, run several rounds of immunization until the title is satisfactory.

Next question also deals with time. What is the typical throughput turnaround time for the functional assay for a new target?

Function assay. Okay, so it really depends because there are a variety of functional assays. Some are pretty pretty fast, some are pretty time-consuming. Say, high-throughput ELISA binding assays usually just cost us, say, 2 to 3 days, you know, it's pretty, it's pretty fast. But in some assays, you know, we need to construct cell lines or spend more time to prepare all materials, it's time-consuming. So usually we will discuss with our clients to make sure that, you know, which are— what essays are they most interested and they want us to perform these essays for them.

Very good. Good communication with the client is critical. Next question, could you comment on developability parameters specifically for the bispecifics? In particular, does the placement of the fragment affect developability?

Okay, that's a very important question, and it's also a very big question. I think To answer that question, we need another several webinars. And yeah, I would say that bispecific antibodies are pretty important, pretty important in the development of antibody drugs. But currently we still face a lot of challenges because Because, you know, actually we haven't had a very thorough understanding of the bispecific antibody. You know, even the first step, say the construction and expression of bispecific antibody, is a big challenge. We often see that, you know, we can hardly obtain the very homogeneous antibody forms we expected. And we often see some unexpected, say, precipitates or poor structures. And yeah, all— I think all these are associated with, you know, the knowledge, you know, that, you know, requires, you know, to achieve better understanding of bispecific antibodies. And say, think, you know, so to me, the most important parameters First of all is stability and the good expression level and purity. And these are usually we came across a lot of troubles with that. You know, if we cannot tackle these, challenges, we cannot, you know, talk about its, say, application potentials or others.

So there aren't really generalizable things to say here. Yep. Okay, sounds good. Thanks. Next question is, can you comment on potential immunogenicity of VHH? Do you do humanization After lead selection, what are the difficulties during humanization that you may face?

Okay, actually I mentioned in my presentation that, uh, nanobodies have high homology to human heavy antibody heavy chain, so it usually think that the, um, it, you know, its immunogenicity is, you know, much better than, you know, mouse-derived antibodies or rabbit-derived antibody. So not all narrow body, you know, requires humanization, but in some case when, you know, the client wants to do the narrow body humanization, is actually more difficult than conventional antibodies because as you can see that the nanobodies are just 1/10 size of the conventional antibodies. So the operation room is pretty small, okay. And usually when we do the humanization with the antibody, we can easily achieve 95% plus, you know, percent humanization. Okay, but as to narrow body, okay, it's very difficult. Usually the humanization percentage is around 80% because, you know, they their sequence is pretty short. But I think, you know, as to the nanobodies, we usually, we don't worry a lot about its immunogenicity. And, okay, let me see. Do we do humanization after leader selection? Okay, it really depends, depends on our clients' objectives and their applications. Yeah, some choose to do the humanization, some will not. Okay, I think the most difficult part in the humanization of narrowbodies is the back mutation, is the back mutation part, you know, because, you know, Usually we don't want to see many back mutations because that will lower down the percentage of humanization. But as to VHH, usually it requires more back mutations, and the selection of back mutation should be very careful and critical for the success of the humanization of VHH.

BHH. Okay, thanks very much. The next question, what is the sample size per experiment on the single B cell sorting platform?

About how many positive cells can be obtained? Okay, that's a very good question. And actually, many people are interested in that new single B cell sorting platform, and we Usually in a typical project, we load 1 million B cells, you know, each one, 1 million. And so I think that's an optical sample size. With that, with 1 million B cells loaded, usually, you know, the encapsulation is pretty good. We seldom see multiple B cells encapsulated in one oil droplet. Okay, so yeah, you can imagine that, you know, if there's a huge amount of B cells loaded and the chance of multiple B cells encapsulated in one tiny oil droplet increased dramatically. So we tried many times and we finally see that 1 million is an optimal sample size. And let me see, oh, how many positive cells? Okay, usually thousands of positive cells can be obtained after each run.

Okay, yes, excellent, excellent. One last question on the single B cell sorting.

Can you perform— Okay, so that's all.

We can go ahead to the next question. Yes, the next question is, can you perform more than one assay in your single B cell sorting? For example, human and Sino antigen binding and blocking, or can you perform multiple rounds of sorting with a different assay? Can you read the last question? Hi, Janice. Can you hear me now? Oh, yes. Yeah. Oh, sorry. Yeah, I'm not sure what happened there. I think we have the one last question. Okay. And that is Can you perform more than one assay in your B cell sorting? For example, human and Sino antigen binding and blocking, or can you perform multiple rounds of sorting with a different assay each time?

Um, so far it's difficult. Is difficult because, you know, you can think about the FRET effect. You know, that's very sensitive and happens in tiny structure. And it's hard, it's very hard, you know, to have double donor and a double receptor system in, you know, formed in that tiny oil droplet, and it will be very complicated. And so far, we don't think it's feasible. So, and what about multiple rounds of sorting with a different assay? Okay, yeah, multiple rounds of sorting is possible. Yeah, it's possible, but usually We think that when one round is enough, then we just go ahead to the sequencing and the bioinformatics analysis to select the candidate sequences. Then we do the expression and the validation, and in the validation we can perform binding assays against the different antigens.

Okay, thanks very much. I think that's all the questions, and just let me double-check here. You can still hear me? Yes, yes, very clearly. Good. I don't know where I went to last time. Thanks very much. In concluding, I would like to thank Dr. Xu for providing insights on innovative technologies and platforms that are used in antibody discovery. I thank you for joining the webinar today, and On-demand version will be available in about a day. I'll send a link to it via everyone who— by email to everyone who registered. Please do feel free to watch this or any of our on-demand webinars when it's convenient.

Thanks again and have a great rest of your day.

Thank you.

Bye.