The lines that start with ">" are comments. In the dataset, comments and sequences with ambiguous amino acid types are removed, resulting in 440115 sequences. Each line contains one protein sequence. Here are the encodings of the amino acid types: Amino acid Three letter code One letter code alanine ala A arginine arg R asparagine asn N aspartic acid asp D *asparagine or aspartic acid asx B (D or N) cysteine cys C glutamic acid glu E glutamine gln Q *glutamine or glutamic acid glx Z (E or Q) glycine gly G histidine his H isoleucine ile I leucine leu L *leucine or isoleucine J (I or L) lysine lys K methionine met M phenylalanine phe F *pyrrolysine O proline pro P serine ser S *selenocysteine sel U (replaced by C) threonine thr T tryptophan trp W tyrosine tyr Y valine val V any X Here are the frequencies of the amino acid types in the dataset (with sequences with any ambiguous type removed). Total amino acids in the dataset: 109,364,666 9865106 A 3479325 C 6092314 D 7193019 E 4225994 F 9585460 G 2903015 H 6110162 I 6547033 K 9798031 L 2550477 M 4579841 N 5078892 P 4129103 Q 5776441 R 6863229 S 6177017 T 7685372 V 1441017 W 3725275 Y Note that the lines with "*" indicate ambiguous or types that should not exist in the sequences. Those have been removed. In case that you like to know the protein sequences and find more information about them, you can find them in pdb_seqres_org.txt by matching the sequences. pdb_seqres_org.txt contains 461,504 sequences with comments (the protein names). Here are the frequencies of the letters in the original dataset: 10014454 A 33 B 2688583 C 6162478 D 7296552 E 4279863 F 9732998 G 2936765 H 6177244 I 6623545 K 9930827 L 2582302 M 4629130 N 2 O 5141747 P 4178337 Q 5850646 R 6941264 S 6244094 T 885554 U 7774077 V 1459853 W 552126 X 3766184 Y 59 Z