Archived Comments for:
Identification, characterization, and comparative genomic distribution of the HERV-K (HML-2) group of human endogenous retroviruses
Incorrect interpretation of previously published data in the paper ¿Identification, characterization, and comparative genomic distribution of the HERV-K (HML-2) group of human endogenous retroviruses¿ written by Ravi P Subramanian, Julia H Wildschutte, Crystal Russo and John M Coffin.
Anton Buzdin, Shemyakin-Ovchinnikov Institute of Bioorganic Chemistry
17 January 2012
Incorrect interpretation of previously published data in the paper ¿Identification, characterization, and comparative genomic distribution of the HERV-K (HML-2) group of human endogenous retroviruses¿ written by Ravi P Subramanian, Julia H Wildschutte, Crystal Russo and John M Coffin.
Anton A. Buzdin, Eugene D. Sverdlov
Shemyakin-Ovchinnikov Institute of Bioorganic Chemistry, Moscow 117997 Miklukho-Maklaya 16/10
In this paper, Subramanian and colleagues describe the results of their systematic bioinformatics screening of the endogenous retroviral group HERV-K (HML-2) representatives that may exist in human genome both in the form of proviral copies or as the solitary LTR sequences [1]. The paper contains important information on the structural features and genomic distribution of this retroviral group. The authors also provided new estimates of the number of the HERV-K(HML-2) members in human DNA, that are more accurate than several recent estimates.
However, in this publication the authors misinterpret some data published before. For example, the authors state:
¿HML-2 LTRs cluster into one of three subgroups based on phylogeny and shared nucleotide features: LTR5Hs, LTR5A, and LTR5B ..[2].¿
This passage includes citation of our paper published in 2003 describing sequence analysis of the human specific HERV-K (HML-2) inserts, which allowed us to identify a family of HERV-K (HML-2) LTRs that included mostly human specific members (in our paper termed ¿HS family¿). The detailed analysis of this family enabled us to subdivide it into two subfamilies: HS-a and HS-b. The subfamily HS-a included only human specific members, whereas some of the subfamily HS-b members could be also found in the chimpanzee genome. The estimated evolutionary ages for these subfamilies were 5,8 and 10,3 million years, respectively. We published the diagnostic nucleotide positions within the consensus sequence that unambiguously distinguish the HS-a and -b subfamilies. When the paper has been already published, the database of eukaryotic repetitive elements Repbase Update [3] was updated with the consensus sequence of the evolutionary recent fraction of the HERV-K (HML-2) LTRs termed ¿LTR5Hs¿, which had many common features with our ¿HS family¿ consensus sequence, but was not identical to it. Later on, Repbase Update has published two other consensus sequences to characterize evolutionary older HERV-K (HML-2) LTRs, termed LTR5A and LTR5B. Importantly, these consensus sequences built for the evolutionary older LTRs had little in common with our consensus sequences termed HS-a and HS-b created for the youngest subset of the HERV-K (HML-2) LTRs. For example, the consensus sequences from the Repbase Update missed any of the above diagnostic nucleotide substitutions characterizing our HS-a and -b subfamilies.
Unfortunately, Subramanian and colleagues appeared to mix up the two classifications mentioned above (communicated in our paper and in Repbase Update). This mistake, to our mind, led to significant misinterpretation our data in the present manuscript that resulted to the following passages:
¿Overall, our classification largely agrees with the previous report defining the three major subgroups [2]; however, the larger sample size of our data set highlighted inconsistencies in the previous classification system¿In a sequence comparison of 1092 HML-2 LTRs, we successfully identified subgroup-specific features which we used to discriminate the LTR5Hs, LTR5A, and LTR5B elements. We would like to note that previously described sequence polymorphisms [2] were not observed among all sequences in any subgroup, likely due to our large sample size¿ The ages of HML-2 proviruses as a function of LTR subgroup were previously estimated by Buzdin et al. based on the intrabranch divergence between individual elements from the subgroup consensus. Their analysis of ~40 LTRs estimated that the LTR5A and 5B subgroups formed around 5.8 and 10.3 million years ago (mya), respectively [14], with 5A originating from 5B. These are fairly recent estimates for these subgroups, given that most LTR5A and 5B proviruses have shared loci among primates whose divergence from humans substantially predates this timeframe. This underestimation is likely due to faulty molecular clock assumptions as well as the use of relatively few proviruses from early sequence builds¿.Consistent with the LTR-based and internal- based phylogenies (Figures 2, 3, and 4) we found the LTR5A and LTR5B proviruses to have formed earlier, around ~20.1 (± 5.4) mya and ~27.9 (± 12.0) mya, respectively¿.We did not observe the sequence polymorphisms within subgroups of our sequences as previously used to define the groups [2], likely due to our much larger sample set.¿
Subramanian et al. compare the data collected in their study for the groups LTR5A and LTR5B(Repbase Update) with our data on the groups HS-a and HS-b. The disagreement in these datasets, to our mind, is not caused by any kind of methodical faults, but is exclusively due to the structural differences between the LTR5A/B vs HS-a/ b groups.
We believe that in this communication it is extremely important to underline the differnce between the HERV-K (HML-2) LTR classifications published in our paper [2], and those presenting in the Repbase Update database in order to avoid any related confusing situations in the future.
1. Subramanian RP, Wildschutte JH, Russo C, Coffin JM: Identification, characterization, and comparative genomic distribution of the HERV-K (HML-2) group of human endogenous retroviruses. Retrovirology, 8:90.
2. Buzdin A, Ustyugova S, Khodosevich K, Mamedov I, Lebedev Y, Hunsmann G, Sverdlov E: Human-specific subfamilies of HERV-K (HML-2) long terminal repeats: three master genes were active simultaneously during branching of hominoid lineages. Genomics 2003, 81:149-156.
3. Jurka J, Kapitonov VV, Pavlicek A, Klonowski P, Kohany O, Walichiewicz J: Repbase Update, a database of eukaryotic repetitive elements. Cytogenet Genome Res 2005, 110:462-467.
Incorrect interpretation of previously published data in the paper ¿Identification, characterization, and comparative genomic distribution of the HERV-K (HML-2) group of human endogenous retroviruses¿ written by Ravi P Subramanian, Julia H Wildschutte, Crystal Russo and John M Coffin.
17 January 2012
Incorrect interpretation of previously published data in the paper ¿Identification, characterization, and comparative genomic distribution of the HERV-K (HML-2) group of human endogenous retroviruses¿ written by Ravi P Subramanian, Julia H Wildschutte, Crystal Russo and John M Coffin.
Anton A. Buzdin, Eugene D. Sverdlov
Shemyakin-Ovchinnikov Institute of Bioorganic Chemistry, Moscow 117997 Miklukho-Maklaya 16/10
In this paper, Subramanian and colleagues describe the results of their systematic bioinformatics screening of the endogenous retroviral group HERV-K (HML-2) representatives that may exist in human genome both in the form of proviral copies or as the solitary LTR sequences [1]. The paper contains important information on the structural features and genomic distribution of this retroviral group. The authors also provided new estimates of the number of the HERV-K(HML-2) members in human DNA, that are more accurate than several recent estimates.
However, in this publication the authors misinterpret some data published before. For example, the authors state:
¿HML-2 LTRs cluster into one of three subgroups based on phylogeny and shared nucleotide features: LTR5Hs, LTR5A, and LTR5B ..[2].¿
This passage includes citation of our paper published in 2003 describing sequence analysis of the human specific HERV-K (HML-2) inserts, which allowed us to identify a family of HERV-K (HML-2) LTRs that included mostly human specific members (in our paper termed ¿HS family¿). The detailed analysis of this family enabled us to subdivide it into two subfamilies: HS-a and HS-b. The subfamily HS-a included only human specific members, whereas some of the subfamily HS-b members could be also found in the chimpanzee genome. The estimated evolutionary ages for these subfamilies were 5,8 and 10,3 million years, respectively. We published the diagnostic nucleotide positions within the consensus sequence that unambiguously distinguish the HS-a and -b subfamilies. When the paper has been already published, the database of eukaryotic repetitive elements Repbase Update [3] was updated with the consensus sequence of the evolutionary recent fraction of the HERV-K (HML-2) LTRs termed ¿LTR5Hs¿, which had many common features with our ¿HS family¿ consensus sequence, but was not identical to it. Later on, Repbase Update has published two other consensus sequences to characterize evolutionary older HERV-K (HML-2) LTRs, termed LTR5A and LTR5B. Importantly, these consensus sequences built for the evolutionary older LTRs had little in common with our consensus sequences termed HS-a and HS-b created for the youngest subset of the HERV-K (HML-2) LTRs. For example, the consensus sequences from the Repbase Update missed any of the above diagnostic nucleotide substitutions characterizing our HS-a and -b subfamilies.
Unfortunately, Subramanian and colleagues appeared to mix up the two classifications mentioned above (communicated in our paper and in Repbase Update). This mistake, to our mind, led to significant misinterpretation our data in the present manuscript that resulted to the following passages:
¿Overall, our classification largely agrees with the previous report defining the three major subgroups [2]; however, the larger sample size of our data set highlighted inconsistencies in the previous classification system¿In a sequence comparison of 1092 HML-2 LTRs, we successfully identified subgroup-specific features which we used to discriminate the LTR5Hs, LTR5A, and LTR5B elements. We would like to note that previously described sequence polymorphisms [2] were not observed among all sequences in any subgroup, likely due to our large sample size¿ The ages of HML-2 proviruses as a function of LTR subgroup were previously estimated by Buzdin et al. based on the intrabranch divergence between individual elements from the subgroup consensus. Their analysis of ~40 LTRs estimated that the LTR5A and 5B subgroups formed around 5.8 and 10.3 million years ago (mya), respectively [14], with 5A originating from 5B. These are fairly recent estimates for these subgroups, given that most LTR5A and 5B proviruses have shared loci among primates whose divergence from humans substantially predates this timeframe. This underestimation is likely due to faulty molecular clock assumptions as well as the use of relatively few proviruses from early sequence builds¿.Consistent with the LTR-based and internal- based phylogenies (Figures 2, 3, and 4) we found the LTR5A and LTR5B proviruses to have formed earlier, around ~20.1 (± 5.4) mya and ~27.9 (± 12.0) mya, respectively¿.We did not observe the sequence polymorphisms within subgroups of our sequences as previously used to define the groups [2], likely due to our much larger sample set.¿
Subramanian et al. compare the data collected in their study for the groups LTR5A and LTR5B(Repbase Update) with our data on the groups HS-a and HS-b. The disagreement in these datasets, to our mind, is not caused by any kind of methodical faults, but is exclusively due to the structural differences between the LTR5A/B vs HS-a/ b groups.
We believe that in this communication it is extremely important to underline the differnce between the HERV-K (HML-2) LTR classifications published in our paper [2], and those presenting in the Repbase Update database in order to avoid any related confusing situations in the future.
1. Subramanian RP, Wildschutte JH, Russo C, Coffin JM: Identification, characterization, and comparative genomic distribution of the HERV-K (HML-2) group of human endogenous retroviruses. Retrovirology, 8:90.
2. Buzdin A, Ustyugova S, Khodosevich K, Mamedov I, Lebedev Y, Hunsmann G, Sverdlov E: Human-specific subfamilies of HERV-K (HML-2) long terminal repeats: three master genes were active simultaneously during branching of hominoid lineages. Genomics 2003, 81:149-156.
3. Jurka J, Kapitonov VV, Pavlicek A, Klonowski P, Kohany O, Walichiewicz J: Repbase Update, a database of eukaryotic repetitive elements. Cytogenet Genome Res 2005, 110:462-467.
Competing interests
No competitive interests declared