Thu Dec 05, 2019 5:48 pm
Re: Cadmium
Hi Jahangir,
For barcoding, I would recommend using a minimum number of channels required to achieve the barcoding level that you need. This is because in my testing (and in most antibody-based barcoding schemes, including the CD298/b2m/CD45 pioneered by Felix), I only used one clone for the CD45 epitope. Therefore, each CD45 barcode reagent will be in competition with each other CD45 barcode reagent. Henrik showed this effect nicely in his first barcoding paper, where going from 1 to 2 to 3 barcode channels per barcode had a stepwise decrease in barcode channel signal intensity (and therefore resolution of the barcode). Even in my own data presented earlier, you can see that 1/6 barcodes are almost 1e3, while 3/6 barcodes are lower at just above 1e2. Additionally, the impurity spillovers are additive the more barcodes you use, thus increasing your background.
For example, 6 BC channels would give you 1/6/15/20/15/6/1 possibilities for 0/1/2/3/4/5/6-of-6 codes (remember, Pascal's triangle is a quick way to calculate this). So, if you need 15 BCs, you can do 2-of-6 or 4-of-6. 2-of-6 would save you on reagents, but also have higher signal in each of the channels you used than would the corresponding 4-of-6. Of course, if you wanted 20 barcodes, you have to do 3-of-6 (assuming 6 BC channels).
As such: 7 BC channels give you 1/7/21/35/35/21/7/1 for 0/1/2/3/4/5/6/7-of-7 codes. If you do 3-of-7 or 4-of-7, you get the same number of BC plexing.
Therefore, I don't see why you would do 4-of-7: you a) don't get any increase in barcoding plexing, b) have more reagent cost, and c) have lower signal (lower resolution due to decrease in Pos and potential increase in Neg) per BC channel. To me, these issues don't sufficiently offset any theoretically minor increase in "purity" (especially the lower resolution issue).
Mike
For barcoding, I would recommend using a minimum number of channels required to achieve the barcoding level that you need. This is because in my testing (and in most antibody-based barcoding schemes, including the CD298/b2m/CD45 pioneered by Felix), I only used one clone for the CD45 epitope. Therefore, each CD45 barcode reagent will be in competition with each other CD45 barcode reagent. Henrik showed this effect nicely in his first barcoding paper, where going from 1 to 2 to 3 barcode channels per barcode had a stepwise decrease in barcode channel signal intensity (and therefore resolution of the barcode). Even in my own data presented earlier, you can see that 1/6 barcodes are almost 1e3, while 3/6 barcodes are lower at just above 1e2. Additionally, the impurity spillovers are additive the more barcodes you use, thus increasing your background.
For example, 6 BC channels would give you 1/6/15/20/15/6/1 possibilities for 0/1/2/3/4/5/6-of-6 codes (remember, Pascal's triangle is a quick way to calculate this). So, if you need 15 BCs, you can do 2-of-6 or 4-of-6. 2-of-6 would save you on reagents, but also have higher signal in each of the channels you used than would the corresponding 4-of-6. Of course, if you wanted 20 barcodes, you have to do 3-of-6 (assuming 6 BC channels).
As such: 7 BC channels give you 1/7/21/35/35/21/7/1 for 0/1/2/3/4/5/6/7-of-7 codes. If you do 3-of-7 or 4-of-7, you get the same number of BC plexing.
Therefore, I don't see why you would do 4-of-7: you a) don't get any increase in barcoding plexing, b) have more reagent cost, and c) have lower signal (lower resolution due to decrease in Pos and potential increase in Neg) per BC channel. To me, these issues don't sufficiently offset any theoretically minor increase in "purity" (especially the lower resolution issue).
Mike