The numbering schemes are rather confusing, particularly in CDR-1 where the equivalents change depending on the length of the CDR.

Anne Marie Honneger's pages show a useful comparison of numbering schemes. However, this comparison doesn't show the equivalent residues in different schemes.

---

CDR-L1

For CDR-L1, the lengths are expressed between residues L24 and L34 (the numbers for these residues do not change between Chothia and Kabat schemes).

Further confusion comes from the fact that Chothia (erroneously) changed his numbering scheme for some paper where the insertion occurred after L31 rather than after L30

Length 10

Kabat	24	25	26	27	29	30	31	32	33	34
Chothia	24	25	26	27	28	29	30	32	33	34

Length 11

Kabat	24	25	26	27	28	29	30	31	32	33	34
Chothia	24	25	26	27	28	29	30	31	32	33	34

Length 12

Kabat	24	25	26	27	27a	28	29	30	31	32	33	34
Chothia	24	25	26	27	28	29	30	30a	31	32	33	34

Length 13

Kabat	24	25	26	27	27a	27b	28	29	30	31	32	33	34
Chothia	24	25	26	27	28	29	30	30a	30b	31	32	33	34

Length 14

Kabat	24	25	26	27	27a	27b	27c	28	29	30	31	32	33	34
Chothia	24	25	26	27	28	29	30	30a	30b	30c	31	32	33	34

Length 15

Kabat	24	25	26	27	27a	27b	27c	27d	28	29	30	31	32	33	34
Chothia	24	25	26	27	28	29	30	30a	30b	30c	30d	31	32	33	34

Length 16

Kabat	24	25	26	27	27a	27b	27c	27d	27e	28	29	30	31	32	33	34
Chothia	24	25	26	27	28	29	30	30a	30b	30c	30d	30e	31	32	33	34

Length 17

Kabat	24	25	26	27	27a	27b	27c	27d	27e	27f	28	29	30	31	32	33	34
Chothia	24	25	26	27	28	29	30	30a	30b	30c	30d	30e	30f	31	32	33	34

---

CDR-L3

The numbering schemes for Kabat and Chothia CDR-L3 are identical. The Kabat CDR runs from L89 to L97.

However, in the Kabat data, there are discrepancies about where the deletion should occur for shorter loops (i.e. length 8). Looking in Kabat at the CDR-L3 loops where residue L97 is present (i.e. not truncated sequences), in 139 cases L95 is present and L96 is missing while in 140 cases L95 is missing and L96 is present. i.e. in the 8-residue loops, Kabat is completely unclear whether the residues should be numbered L89-L90-L91-L92-L93-L94-L96-L97, or whether they should be numbered L89-L90-L91-L92-L93-L94-L95-L97.

While a number of PDB files including 1yqv and 1a7o delete L96 from the numbering, it is clear on comparing the structures of 8 and 9 residue CDR-L3 loops, that it should be L95 that is removed.

The correct numbering is therefore:

Length 8

Kabat/Chothia	89	90	91	92	93	94	96	97

Length 9

Kabat/Chothia	89	90	91	92	93	94	95	96	97

Length 10

Kabat/Chothia	89	90	91	92	93	94	95	95A	96	97

Length 11

Kabat/Chothia	89	90	91	92	93	94	95	95A	9B	96	97

...etc...

---

CDR-H1

For CDR-H1, the situation is even more confusing because, in the Kabat numbering scheme, the insertion is at the end of the loop. Again, the equivalences depend on the length of the loop, so the lengths shown below are for residues H41-H45B (Kabat numbering) or H41-H45 (Chothia numbering).

Length 5

Kabat	31	32	33	34	35
Chothia	31	32	33	34	35

Length 6

Kabat	31	32	33	34	35	35a
Chothia	31	31a	32	33	34	35

Length 7

Kabat	31	32	33	34	35	35a	35b
Chothia	31	31a	31b	32	33	34	35