CN114242174B - Identification and annotation method for endogenous retroviruses - Google Patents

Abstract

The invention provides an identification annotation method for endogenous retroviruses, which comprises the following steps: selecting virus protein as a probe, identifying a hit segment similar to the probe, outputting a hit region, and extending flanking sequences to two sides of the hit region to obtain an endogenous retrovirus candidate sequence; identifying paired LTR sequences from endogenous retrovirus candidate sequences by using LTR harvest, and extracting potential complete endogenous retrovirus sequences and endogenous retrovirus candidate sequences without paired LTR sequences; identifying a virus protein structural domain by using a hidden Markov model based on a typical protein structural domain sequence of the retrovirus, and removing a false positive result; endogenous retroviruses were annotated and protein domain sequences were extracted. The method can realize the rapid and efficient excavation, identification and annotation of endogenous retroviruses and elements of the host genome, and greatly reduce the false positive rate.

Description

A method for identification and annotation of endogenous retroviruses

technical field

The invention relates to the technical field of gene detection, in particular to an identification and annotation method for endogenous retroviruses.

Background technique

The viral components in the host genome are called endogenous retroviruses (ERVs). When retroviruses infect the host, they can integrate their own genome or some elements into the host genome through some steps in the replication process. . Many endogenous retroviruses have been integrated into the host genome millions of years ago, and as the ancient "fossil record" of the co-evolution of viruses and life (host), they provide raw material for studying the origin and evolution of life. During long-term co-evolution, they have become part of the host genome, and some of them have even evolved into functional genes that play important biological functions, such as the syncytin gene associated with embryonic development, derived from endogenous reversal The viral envelope protein. In addition, endogenous retroviruses and their elements integrated into the genome can also exert antiviral functions directly or indirectly. However, most of the endogenous retroviruses or elements integrated into the host genome are destroyed because of their integrity, and the genetic variation events during the long evolution process often have low homology with modern exogenous viruses. Therefore, , whose identification and annotation has always been a difficulty. In general, there is still a lack of efficient identification and annotation methods for endogenous retroviruses and elements in the host genome. CN201710418223.3 discloses a method for identifying and breeding a new line of PERV-pol gene-deficient Wuzhishan miniature pigs, using PCR and/or RT-PCR to identify whether the cells carry the structure of porcine endogenous retrovirus Gene, and then can identify whether the pig to be tested is an inbred line of Wuzhishan miniature pigs deficient in endogenous retrovirus pol gene. This method has low efficiency and shortcomings in the actual application process.

Based on this, the present invention proposes a method for identifying and annotating endogenous retroviruses, which can quickly and efficiently perform mining, identification and annotation of endogenous retroviruses and elements in the host genome.

SUMMARY OF THE INVENTION

The purpose of the present invention is to provide a method for identifying and annotating endogenous retroviruses, which combines homologous sequence search and hidden Markov model prediction method to realize rapid and efficient identification and annotation of endogenous retroviruses.

Based on above-mentioned technical purpose, the present invention adopts following technical scheme:

In one aspect, the present invention provides a method for identifying and annotating endogenous retroviruses, the specific steps comprising:

Step 1): Select a typical and relatively conservative viral protein as a probe, use the homologous sequence alignment method to identify the hit fragments similar to the probe, output a reasonable hit region for the consecutive hit fragments, and send the reasonable hit to the reasonable hit. The extended flanking sequences on both sides of the region, after removing the region fragments containing the inclusion relationship, obtain the endogenous retrovirus candidate sequence;

Step 2): Use LTR harvest to identify the paired long terminal repeat sequences of the endogenous retroviral candidate sequences obtained in step 1), and then extract the potential complete endogenous retrovirus sequences and those that do not contain paired LTR sequences. endogenous retroviral selection sequences;

Step 3): Potential complete endogenous retrovirus sequences extracted from step 2) and endogenous retroviruses that do not contain paired LTR sequences were identified using a hidden Markov model based on the canonical protein domain sequences of retroviruses The viral protein domain of the selected sequence is predicted based on the hidden Markov model, and the false positive results generated by the homologous sequence alignment method in the previous step 1) are removed;

Step 4): Based on the viral protein domain identified in step 3), locate the host genome, generate an annotation file on the host genome, extract the full-length sequence and viral domain sequence of the endogenous retrovirus, and generate the corresponding Locate and annotate files.

Further, the typical and relatively conserved viral protein is selected from at least one of: Pol protein sequence, ENV protein sequence, and RT protein sequence.

Further, the homologous sequence alignment method is selected from the BLAST local alignment method.

Further, the specific process of the output of the hit region is as follows: the frameshift event of the viral sequence in the host genome will lead to the occurrence of many consecutive hit fragments, and by judging whether the distance between each hit fragment is reasonable (the reasonable standard for judging is greater than 10kb). length bases), whether the insertion direction on the host genome is the same (the same is classified as a contiguous hit fragment on the same strand of the genome, and if it is not the same, it is classified as a contiguous hit fragment on a different chain of the genome), and the identification may be over-shifted. The mutated endogenous retroviral fragment outputs a reasonable hit region for consecutive hits on each genome strand, and returns the location of the hit region on the genome.

Further, the length of the flanking sequence is at least 15kb, that is, the above-mentioned reasonable hit region extends the base sequence of at least 15kb length to both sides, after removing the region fragments containing the inclusion relationship, then extract these fragment sequences as endogenous. Sexual retroviral candidate sequences.

Further, in step 2), the specific method for identifying long terminal repeat sequences is: constructing an enhanced suffix index array for the endogenous retrovirus candidate sequences obtained in step 1), and performing paired LTR sequence search to achieve paired LTR sequences. Identification of long terminal repeats.

Further, in step 2), the search criteria for the paired LTR sequences are: only LTR sequences with suitable distance (such as within the range of 1kb to 15kb), suitable sequence length (such as greater than 100bp) and similarity greater than 80% will be used. Considered candidate paired LTR sequences.

Further, the extraction process of the potential complete endogenous retrovirus sequence and the endogenous retrovirus selection sequence that does not contain the paired LTR sequence is: for each candidate paired LTR sequence and the region in the middle, Take them as potential complete endogenous retroviral sequences, return their positions on the genome, remove the repetitive regions according to the positioning, and then extract these potential complete endogenous retroviral sequences; Endogenous retrovirus-selected sequences for LTR sequences were also extracted separately and used for identification of incomplete endogenous retroviruses and their viral elements.

Further, the retrovirus protein domain sequence refers to: download the typical protein domain sequence of retrovirus from a retrovirus protein structure related database (such as Gypsy database), including GAG, DUT, PR, RT, INT, RNaseH, ENV, etc.; based on the 6 coding translation frames translated into protein sequences and used as input sequences, the hidden Markov model was used to identify potential complete endogenous retroviral sequences containing paired LTRs, and those without paired LTRs, respectively. Viral protein domains of endogenous retroviral candidate sequences.

Further, in the step 4), based on the input sequence identified in step 3) that contains at least one of the above viral protein domains, (1) if the input sequence is a potentially complete endogenous retrovirus containing paired LTRs Sequence, first locate the LTRs on both sides and the identified viral domains on the host genome, generate an annotation file on the host genome, and then extract the complete endogenous retrovirus full-length sequence and viral domain sequence , flank LTR sequences, and generate the location and annotation files of flank LTRs and viral domains on the complete endogenous retrovirus full-length sequence; (2) If the input sequence is an endogenous reverse without paired LTRs The candidate sequence of the retrovirus, first locate the identified domain on the host genome, generate an annotation file on the host genome, and then extract the potential endogenous retrovirus full-length sequence according to the location on the host genome and viral domain sequences, and generate a mapping and annotation file of viral domains on potential endogenous retroviral full-length sequences.

Compared with the prior art, the technical solutions provided by the present invention have the following beneficial effects: the method of the present invention can rapidly and efficiently perform full excavation, identification and annotation of endogenous retroviruses and elements on the host genome, wherein homologous The combination of sequence alignment and hidden Markov model prediction greatly reduces the false positive rate.

1) The present invention provides a large amount of raw materials for studying the origin and genetic evolution of viruses and life through the efficient identification and annotation of endogenous retroviruses and elements in the host genome.

2) The method of the present invention helps to identify and screen endogenous retroviruses and elements that can directly or indirectly exert antiviral functions, and will promote the vigorous development of related disciplines and fields.

Description of drawings

Figure 1 is a flow chart of the identification and annotation method of endogenous retroviruses.

Detailed ways

The following examples are intended to illustrate the content of the present invention, rather than to further limit the protection scope of the present invention.

Example 1

The present invention prepares a method for identifying and annotating endogenous retroviruses, and the specific method is as follows:

1. Search for endogenous retroviral candidate sequences

Mainly based on the homologous sequence search method, the endogenous retrovirus candidate sequences were preliminarily identified. First select typical and conserved viral proteins as probes (such as using retrovirus Pol protein sequence or ENV protein sequence), and then use homologous sequence alignment methods (such as BLAST local alignment search) to identify the genome with probes Similar hit fragments. If multiple viral proteins are used as probe sequences, the search results for multiple probe sequences are combined. The hits are then sorted according to the same strand (the gene orientation of the hit is the same as the input sequence) and the complementary strand (the gene orientation of the hit is opposite to the input sequence).

The frameshift event of the viral sequence in the host genome will lead to many consecutive hit fragments. By judging whether the distance between the hit fragments is reasonable (the reasonable standard is greater than 10kb length bases), and whether the insertion direction on the host genome is the same (The same is classified as a contiguous hit fragment on the same strand of the genome, and if it is not the same, it is classified as a contiguous hit fragment on a different chain of the genome), identify endogenous retroviral fragments that may have undergone frameshift mutation, and identify consecutive The hit fragment outputs a plausible hit region and returns the location of the plausible hit region on the genome. Finally, for each reasonable hit region, the flanking sequences (flanking sequences such as base sequences with a length of 15 kb) are respectively extended to both sides, and the region fragments containing the inclusion relationship are removed to obtain the endogenous retrovirus candidate sequence.

2. Pairwise long terminal repeat identification

An important feature of intact retroviruses is the presence of long terminal repeats (LTR sequences) flanking the virus. Use LTR harvest to identify long terminal repeat sequences. First, build an enhanced suffix index array for the above-mentioned endogenous retrovirus candidate sequences (the construction of the retrieval index in this step is mainly to speed up the subsequent retrieval speed). This step can save a lot of subsequent calculations. time, followed by a search for paired LTR sequences. Only LTR sequences with suitable distance (such as in the range of 1kb to 15kb), suitable sequence length (such as more than 100bp) and more than 80% similarity will be considered as candidate paired LTR sequences. For each candidate paired LTR sequence and the region in between, use them as a potential complete endogenous retroviral sequence, return its location on the genome, remove the duplicated region according to the location, and then extract these potential The sequence of the complete endogenous retrovirus.

At the same time, the selected sequences of endogenous retroviruses that do not contain paired LTR sequences are also extracted separately and used for the identification of incomplete endogenous retroviruses and their viral elements.

Identification of viral protein domains

Download the typical protein domain sequences of retroviruses, including GAG, DUT, PR, RT, INT, RNaseH, ENV, etc., from databases related to retrovirus protein structure (such as Gypsy database). For each potential complete endogenous retroviral sequence containing paired LTRs, and endogenous retroviral candidate sequences without paired LTRs, based on the 6 coding translation frames, it was translated into 6 protein sequences and used as input sequences, The viral domains were identified using a hidden Markov model, respectively, and the results were predicted based on the hidden Markov model, removing the false positive results previously generated by the homologous sequence alignment method.

Annotation of endogenous retroviruses

For the identified input sequence that contains at least one of the above viral protein domains, (1) if the input sequence is a potentially complete endogenous retroviral sequence containing paired LTRs, first combine the flanking LTRs, the identified viral structure The domain is located on the host genome, an annotation file is generated on the host genome, and the complete endogenous retrovirus full-length sequence, viral domain sequence, and flank LTR sequences are extracted, and the two sides LTR and virus are generated. The localization and annotation files of the domains on the complete endogenous retrovirus full-length sequence; (2) If the input sequence is an endogenous retrovirus candidate sequence without paired LTRs, first place the identified domains in the Positioning on the host genome, generating annotation files on the host genome, and then extracting potential endogenous retrovirus full-length sequences and viral domain sequences based on the positioning on the host genome, and generating viral domains in potential Mapping and annotation files on the full-length sequences of endogenous retroviruses.

Example use:

1. Host genome data download

Download the genome sequence file of the bat species Pteronotus parnellii (GCA_000465405.1) from the NCBI database (https://www.ncbi.nlm.nih.gov/), the genome file size is about 2Gb.

Obtaining endogenous retroviral candidate sequences

(1) Build the above bat genome sequence into a local nucleotide database, and use RT protein and ENV protein sequences as probe sequences. Based on the tblastn method in BLAST, identify the hit fragments on the bat genome that are similar to the probe sequences in the database. .

(2) Combine the BLAST search results of the two probe sequences, and classify the hit regions according to the same strand (the gene direction of the hit region is the same as the input sequence) and the complementary strand (the gene direction of the hit region is opposite to the input sequence). By judging whether the distance between the hit fragments is greater than 10kb, combined with the insertion direction of the hit fragments on the host genome, a reasonable hit region is output for the consecutive hit fragments, and obtained.

(3) For each of the above reasonable hit regions, extend 15kb flanking sequences to both sides, remove the region fragments containing the inclusion relationship, and obtain endogenous retrovirus candidate sequences. A total of 628 endogenous retroviral candidate sequences were identified in the bat species Pteronotus parnellii .

Identification of paired LTRs in endogenous retroviral candidate sequences

Use LTR harvest to construct a retrieval index array for the above 628 endogenous retrovirus candidate sequences, and identify whether each candidate sequence contains paired LTR sequences. The criteria for paired LTR sequences are: the LTR sequence is greater than 100bp, the distance between the two LTR sequences is in the range of 1kb to 15kb, and the similarity between the two LTR sequences is greater than 80%. For each candidate paired LTR sequence and the region in between, use them as a potential complete endogenous retroviral sequence, return its location on the genome, remove the duplicated region according to the location, and then extract these potential The sequence of the complete endogenous retrovirus. Among the above 628 endogenous retrovirus candidate sequences, paired LTR sequences were identified in 38 sequences (as shown in Table 1). The inclusion of "same" in the sequence name indicates that the identified endogenous retrovirus is on the same strand as the input sequence, and the inclusion of "complement" in the sequence name indicates that the identified endogenous retrovirus is on the complementary strand of the input sequence.

Table 1

4. Identification of viral protein domains

Download the typical protein domain sequences of retroviruses from the Gypsy database, including GAG, DUT, PR, RT, INT, RNaseH, ENV, etc. The above 628 endogenous retrovirus candidate sequences were translated into protein sequences based on 6 coding translation frames (3768 in total). For these 3768 protein sequences, viral domains were identified based on the Hidden Markov Model using the above-mentioned typical protein domains of retroviruses.

5. Annotation of viral protein domains

For the 38 endogenous retroviral candidate sequences that contain paired LTRs, and the remaining endogenous retroviral candidate sequences that do not contain paired LTRs, the search results based on the hidden Markov model, if it contains 1 The typical protein domains of retroviruses from one species or more were identified as the final endogenous retrovirus candidate sequences, which could remove some of the false-positive results previously identified by the BLAST method. A total of 25 endogenous retroviral sequences containing paired LTRs were finally identified, and 580 endogenous retroviral sequences or elements without paired LTRs were identified. Finally, the viral protein domains of these candidate sequences are precisely mapped to the host genome and the extracted endogenous retroviruses to generate detailed annotation documents. These endogenous retroviral protein domains are used to identify and screen endogenous retroviruses and elements that can directly or indirectly exert antiviral functions.

Table 2 is an example of an annotated document on the extracted endogenous retrovirus of endogenous retroviral sequences containing paired LTRs. The annotation document details the position of each endogenous retrovirus in the original host genome (corresponding to the number in the sequence name) and which strand it is located on (“same” in the sequence name means it is located on the same strand as the input sequence. , "complement" indicates the start and end positions of each protein domain on the extracted endogenous retrovirus.

Table 2

paired_LTR ERVs 5'-LTR 3'-LTR GAG DUT AP RT RNaseH INT ENV KE836813.1|5550|11251|complement 1..497 5181..5702 1923..2435 2834..2968 3188..3280 3634..4272 not found not found not found AWGZ01387580.1|212|8442|complement 1..274 7979..8231 1756..2202 2776..3015 3103..3342 3754..4167 4950..5090 5127..5546 7350..7487 KE834585.1|28634|33361|complement 1..110 4621..4728 1948..2169 2536..2802 2887..3132 3451..4020 not found not found not found KE910746.1|37935|44952|same 1..306 6692..7018 not found not found not found 2493..2795 3556..3726 3903..4289 not found KE902113.1|5948|11935|complement 1..251 5724..5988 not found not found not found 2616..2912 not found 4185..4307 not found KE839960.1|10792|18188|same 1..531 6942..7397 2897..3199 not found 3855..3917 4165..4533 not found not found 6397..6669 KE870169.1|1525|10645|same 1..230 8873..9121 1877..2023 not found 2568..2654 2933..3631 4365..4634 4860..5192 not found AWGZ01396048.1|10776|19052|complement 1..224 8058..8277 1511..1705 not found 2363..2545 2868..3473 4141..4521 5033..5302 7432..7557 KE827911.1|26386|33260|complement 1..216 6644..6875 not found not found not found 2315..2635 not found 4402..4527 not found KE893677.1|2553|9228|same 1..235 6460..6676 1919..2047 not found 2663..2743 3015..3257 4306..4608 5103..5312 not found KE842560.1|4542|9092|complement 1..146 4422..4551 not found not found not found 971..1207 not found not found not found KE834830.1|23532|26760|same 1..126 3096..3229 not found not found not found not found not found 440..1342 not found KE873916.1|24431|30941|same 1..163 6350..6511 not found not found not found 2514..2639 not found 4383..4541 5926..6054 AWGZ01271927.1|5500|7412|complement 1..200 1715..1913 not found not found not found 289..693 1687..1905 not found not found KE824373.1|19456|27134|complement 1..201 7449..7679 not found not found 2504..2617 2895..3074 4211..4615 5194..5280 not found AWGZ01389043.1|13908|22015|complement 1..225 7874..8108 2023..2181 not found 2561..2767 3078..3416 4503..4790 5020..5409 7072..7677 KE839612.1|16878|25796|same 1..134 8788..8919 2810..3007 not found not found 4115..4549 5296..5520 6139..6318 8086..8421 KE919690.1|3176|9835|complement 1..194 6479..6660 not found not found 3048..3242 3718..3942 4792..5097 5581..5991 not found KE849596.1|25855|31240|complement 1..186 5194..5386 1897..1998 not found 2599..2667 2985..3185 4243..4431 not found 4740..4844 KE874998.1|3250|13590|complement 1..139 10210..10341 3800..4087 not found not found 5251..5724 6498..6749 7129..7500 9172..9384 KE820627.1|23|7334|same 1..454 6842..7312 3312..3425 not found not found 4132..4341 not found not found 6454..6540 KE905742.1|4419|10795|complement 1..138 6249..6377 1133..1333 not found not found 2733..2921 3641..3907 4302..4532 not found AWGZ01282780.1|7302|14150|same 1..194 6661..6849 not found not found not found 2521..2841 not found 4606..4815 6126..6440 AWGZ01040320.1|700|1872|complement 1..113 1061..1173 not found not found not found not found not found not found 785..922 KE887893.1|13652|18794|same 1..361 4785..5143 not found not found not found not found not found not found 4417..4524

Although the present invention has been described in detail above with general description, specific embodiments and tests, some modifications or improvements can be made on the basis of the present invention, which is obvious to those skilled in the art . Therefore, these modifications or improvements made without departing from the spirit of the present invention fall within the scope of the claimed protection of the present invention.

Claims (10)

1. a kind of identification annotation method for endogenous retrovirus, is characterized in that, concrete steps comprise: Step 1): Select a typical and relatively conservative viral protein as a probe, use the homologous sequence alignment method to identify the hit fragments similar to the probe, output a reasonable hit region for the consecutive hit fragments, and send the reasonable hit to the reasonable hit. The extended flanking sequences on both sides of the region, after removing the region fragments containing the inclusion relationship, obtain the endogenous retrovirus candidate sequence; Step 2): Use LTR harvest to identify the paired long terminal repeats of the endogenous retroviral candidate sequences obtained in step 1), and then extract the potential complete endogenous retrovirus sequences and those that do not contain paired LTR sequences. endogenous retroviral selection sequences; Step 3): Potential complete endogenous retrovirus sequences extracted from step 2) and endogenous retroviruses that do not contain paired LTR sequences were identified using a hidden Markov model based on the canonical protein domain sequences of retroviruses The viral protein domain of the selected sequence, and further remove the false positive result based on the homologous sequence alignment method in step 1); Step 4): Based on the viral protein domain identified in step 3), locate the host genome, generate an annotation file on the host genome, extract the full-length sequence and viral domain sequence of the endogenous retrovirus, and generate the corresponding Mapping and annotation files on endogenous retroviruses. 2. The method for identifying and annotating endogenous retroviruses according to claim 1, wherein the typical and relatively conservative viral proteins are selected from the group consisting of: Pol protein sequence, ENV protein sequence, RT protein sequence. at least one. 3. The method for identifying and annotating endogenous retroviruses according to claim 1, wherein the homologous sequence alignment method is selected from the BLAST local alignment method. 4. according to the method for identifying and annotating endogenous retroviruses according to claim 1, it is characterized in that, the specific process of outputting the hit region is as follows: the frameshift event of the virus sequence in the host genome can cause many serializations to occur. By judging the distance between each hit fragment and the insertion direction on the host genome, identify endogenous retrovirus fragments that may have undergone frameshift mutation, output a reasonable hit region for consecutive hit fragments, and return The location of the hit region on the genome. 5. The method for identifying and annotating endogenous retroviruses according to claim 1, wherein the length of the flanking sequence is at least 15kb. 6. The method for identifying and annotating endogenous retroviruses according to claim 1, characterized in that, in step 2), the specific method for identifying long terminal repeats is: reversing the endogenous retrovirus obtained in step 1). An enhanced suffix index array is constructed from the candidate sequences of the record virus, and the paired LTR sequence search is performed to realize the identification of paired long terminal repeats. 7. The method for identifying and annotating endogenous retroviruses according to claim 1, wherein in step 2), the search criteria for the paired LTR sequences are: a distance of 1kb to 15kb, a sequence length greater than LTR sequences with 100 bp and more than 80% similarity are candidate paired LTR sequences. 8. The method for identifying and annotating endogenous retroviruses according to claim 1, wherein the potentially complete endogenous retrovirus sequences and the endogenous retroviruses that do not contain paired LTR sequences The selected sequence extraction process is as follows: for each candidate paired LTR sequence and the region in between, take them as potential complete endogenous retroviral sequences, return their location on the genome, and then remove duplicates according to the location. Matching regions, and then extract these potentially complete endogenous retroviral sequences; at the same time, for endogenous retroviral selected sequences that do not contain paired LTR sequences, they are also extracted separately and used as incomplete endogenous retroviruses Identification of the virus and its viral elements. 9. The method for identifying and annotating endogenous retroviruses according to claim 1, wherein the retrovirus protein domain sequence refers to: downloading retroviruses from a retrovirus protein structure correlation database. The canonical protein domain sequences, based on the 6 coding translation frames, were translated into protein sequences and used as input sequences to identify potential complete endogenous retroviral sequences with paired LTRs and without paired LTRs using Hidden Markov Models, respectively The viral protein domains of endogenous retroviral candidate sequences. 10. The method for identifying and annotating endogenous retroviruses according to claim 1, wherein in the step 4), based on the input identified in step 3) that at least contains one of the above-mentioned viral protein domains Sequence, (1) If the input sequence is a potential complete endogenous retroviral sequence containing paired LTRs, first locate the LTRs on both sides and the identified viral domains on the host genome, and generate a sequence on the host genome. Annotation file, and then extract the complete endogenous retrovirus full-length sequence, viral domain sequence, and flanking LTR sequences, and generate the flanking LTR and viral domains on the complete endogenous retrovirus full-length sequence. Location and annotation file; (2) If the input sequence is an endogenous retrovirus candidate sequence without paired LTR, first locate the identified domain on the host genome to generate an annotation file on the host genome , and then extract the potential endogenous retrovirus full-length sequence and viral domain sequence according to the location on the host genome, and generate the location and annotation of the viral domain on the potential endogenous retrovirus full-length sequence document.

Images