FEATURES/Location/Qualifiers で記述できないその他の情報やコメントなどが記載されています。
例えば、登録者の所属が複数ある場合、REFERENCE 1 で記載されていない方を記載することがあります。

例
COMMENT     Human cDNA sequencing project.

サンプル採取に関する詳細を "structured COMMENT" に記載

サンプル採取に関する詳細な情報を記載する目的で [項目名] と [項目の値] の組で構造化された COMMENT 行を記載することが可能です。この記法は生物多様性に関する登録, ゲノム全長規模の登録で推奨されています。

例
COMMENT     ##Genome-Assembly-Data-START##
            Finishing Goal           :: Finished
            Current Finishing Status :: High Quality Draft
            Assembly Method          :: Newbler v. 2.3
            Genome Coverage          :: 30x
            Sequencing Technology    :: 454/Illumina
            ##Genome-Assembly-Data-END##

MGA データの生成手法

MGA データには、登録配列が生成されるまでの過程（シーケンス用サンプルの調製法、生の配列データから登録配列への処理方法など）が記載されています。

例
COMMENT     The CAGE (cap analysis gene expression) is based on preparation
            and sequencing of concatamers of DNA tags deriving from the
            initial 20/21 nucleotides from 5' end mRNAs.
            Full-length cDNAs were at first selected with the Cap-Trapper
            method. Then, a specific linker (Linker1, some linker contain 5 bp
            sequences that have 15 variations for each rna sample) containing
            the ClassIIs restriction enzyme site MmeI was then ligated to the
            single-strand cDNA and then the second strand of cDNA synthesized.
            The resulting double-stranded cDNA was cleaved by the restriction
            enzyme MmeI and a second linker (Linker2) was ligated to the 2 bp
            overhang at the MmeI cleaved site, to produce a 5' 20/21 tag
            having two linkers at both sides. The ligation products were
            separated from unmodified DNA with magnetic beads. The 5' end cDNA
            tags were released from the beads, and the DNA fragments were
            amplified in a PCR step by using the two linker-specific primers
            (Primer1 (uni-PCR), Primer2 (MmeI-PCR)). The desired 32-37 bp tags
            were purified and ligated to form concatamers, and then the
            concatamer were fractionated and ligated to the plasmid ZErO-2.
            The ligations were finally electroporated into DH10b cells
            (Invitrogen) and obtained plasmids were sequenced with forward
            primers.
            CAGE libraries were sequenced with forward primers essentially as
            described with minor modifications to use zeocin for selection of
            recombinants. We used in-house developed algorithms for the
            extraction of tags and for masking the vectors. CAGE tags were
            extracted with the following parameters: vector masking, minimum
            12 bp recognition allowed; linker (13 bp) masking: maximum
            mismatch, 2 bp allowed; XmaJI site maximum mismatch, 2 bp allowed;
            tag length, 17-24 bp.
            Linker1: "Upper oligonucleotide GN6":
            biotin-agagagagacctcgagtaactataacggtcctaaggtagcgacctagg (5 bp)
            tccgacGNNNNN and "Upper oligonucleotide N6":