Multiple Most possible strains

[shiny1110]

Hello, after building custom database and performing strain identification on my samples, I noticed that there are multiple "Most possible strains" with identical Total_Map_Rate and Strain_Depth values. Does this indicate the presence of multiple dominant strains in the sample? Thank you!

[liaoherui]

Hi, thanks for using VirStrain!

"Does this indicate the presence of multiple dominant strains in the sample?" — This indicates a tie case, meaning the correct strain is likely one of the listed strains. Tie cases often occur in low-depth datasets. For more details, please refer to our manuscript.

If you have any other questions, feel free to let me know!

[shiny1110]

Thank you very much for your previous response. After carefully reading the manuscript, I have another question regarding strain abundance calculation:
When multiple strains in both "Most possible strains" and "Other possible strains" share identical Valid_Map_Rate and Strain_Depth, how can I calculate the relative abundance of co-existing strains in such scenarios? Should I select one strain from those with identical metrics and calculate their relative abundance based on Strain_Depth?

[liaoherui]

Yes, that's one potential solution. Additionally, you might consider using the -s parameter (manual coverage function, see the GitHub README) when building the database to include more informative sites, which could help reduce tie cases.

[shiny1110]

Hello, I'm currently using a database built with 219 complete viral sequences, where the -s parameter was set to the maximum recommended value of 0.6 during database construction. If I further increase the -s parameter beyond 0.6, what potential impacts might this have on the tool's performance or strain identification accuracy? Thank you for your insights!

[liaoherui]

Hi, we haven’t formally tested the impact of the -s parameter yet. However, in theory (and my guess), using more sites could improve the tool’s resolution by reducing tie cases—since more features may help distinguish the correct strain from others. However, some of these additional sites could be noisy, so setting a very large -s value might increase the risk of false positives.

[shiny1110]

Thank you very much for your detailed response. I have a few additional questions I’d like to ask for clarification:
What do the entries listed under >> Other possible strains represent? For example, an entry like:'>> (Could be FP) 1240-C:4889, 4849-A:1887, 1639-C:504, 1559-C:393, Genome_num: 6'. Does ‘Could be FP’ stand for ‘false positive’? Do these values (e.g., 1240-C:4889) refer to specific mutations, genomic positions, or other metrics?
If the Valid_Map_Rate for entries under Other possible strains is less than 10/​**​*, does this imply that ​​only​​ the "Most possible strains" are present in the sample?

[liaoherui]

Hi, sorry for the late reply.

"Could FP stand for 'false positive'?" → Yes. This means these strains may be false positives because they cover the same number of sites as the most probable strain but have a lower identification score.
"Do these values (e.g., 1240-C:4889)..." → This value refers to the informative sites selected by the program. 1240 refers to the position in the multiple sequence alignment, C refers to the nucleotide, and 4889 refers to the k-mer frequency for this position.
"If the Valid_Map_Rate for entries under Other possible..." → It's hard to say. "Valid_Map_Rate" refers to how many sites owned by this strain are not owned by the most probable strain. So, if there are two similar strains co-existing in the sample, even if this value is quite low, it could still represent a true strain.

[shiny1110]

Thank you once again for your patient and detailed response! I would like to follow up with another question regarding my analysis using VirStrain. After processing my data, I obtained the following results:
Strain_ID Cls_info SubCls_info Map_Score Valid_Map_Rate Total_Map_Rate Strain_Depth Strain_info Unique_SNP

Most possible strains:
>MW265929.1 Cluster85_1 NA 0.9450348662750748 1539/2006 1539/2006 52618.25913891145 {'997-T': 54581...}
Other possible strains:
>>6369-T:71766,1777-C:70982,1785-G:70915,1834-T:69714,8241-G:69070,6141-A:68737,1871-C:68655,1802-G:68576,1816-G:65618,1736-T:65473,3169-T:63937,2083-A:63842,2915-C:63702,4711-T:63211,6195-T:62761,1747-C:62343,4225-C:62280,1759-T:62219,3001-T:62193,5972-G:61902,4762-A:61412,2068-C:61372,6051-T:59772,2677-C:59567,2107-C:58497,1717-G:57799,3796-G:57474,2143-T:57150,3184-G:56849,2615-C:55685,1512-T:55353,2704-T:55096,2617-A:55011,4252-T:54970,2263-A:54756,3118-G:54694,6180-G:54230,3817-G:54112,3352-G:53879,2847-A:53810,2761-C:53797,4033-G:53522,3196-C:52816,2752-A:52794,5025-A:52684,2554-C:52600,5030-G:52554,1648-C:52496,2758-G:51935,5253-C:51687,2779-C:51628,4019-T:51546,3718-T:51417,4015-C:50403,3715-G:49726,3832-T:49499,5275-A:48843,4627-T:48815,4639-C:48625,2458-T:48553,2500-T:48385,4576-A:47844,4635-A:47829,4633-C:47794,3217-G:47292,2338-C:47029,4285-A:46850,1483-C:46211,1481-C:46189,3860-G:45827,4819-A:45679,6021-C:45088,4297-T:45045,2386-A:44975,1274-A:44064,2389-G:42972,3901-A:42899,3256-C:42138,5835-C:42040,1135-T:42033,1134-C:41933,3967-A:40828,1204-C:40410,3694-G:39726,4378-T:35888,1231-T:35824,1222-T:34715,3448-C:30647,3445-G:29399,Genome_num: 1
>MF036686.1 Cluster96_1 NA 0.941579798184874 57/2007 1535/2007 52985.64444444444 Not_record
>>1834-T:69714,8241-G:69070,1816-G:65618,2083-A:63842,6195-T:62761,2068-C:61372,2677-C:59567,2107-C:58497,3184-G:56849,2615-C:55685,1512-T:55353,2617-A:55011,2263-A:54756,6180-G:54230,3817-G:54112,2761-C:53797,3196-C:52816,2752-A:52794,5025-A:52684,2554-C:52600,1648-C:52496,2779-C:51628,2458-T:48553,1483-C:46211,1481-C:46189,2386-A:44975,1274-A:44064,3256-C:42138,5835-C:42040,1135-T:42033,1204-C:40410,3694-G:39726,Genome_num: 1
>PP437897.1 Cluster107_1 NA 0.6073279800662044 15/2007 988/2007 53495.818181818184 Not_record
>>8241-G:69070,1816-G:65618,2677-C:59567,2107-C:58497,3184-G:56849,2615-C:55685,1512-T:55353,6180-G:54230,3817-G:54112,2761-C:53797,2779-C:51628,2458-T:48553,1483-C:46211,3256-C:42138,5835-C:42040,1204-C:40410,3694-G:39726,Genome_num: 1
>MF770715.1 Cluster98_1 NA 0.9398825989676968 7/2007 1526/2007 53354.4 Not_record
>>8241-G:69070,1816-G:65618,2107-C:58497,1512-T:55353,3817-G:54112,2779-C:51628,2458-T:48553,3256-C:42138,1204-C:40410,3694-G:39726,Genome_num: 1
>PP437947.1 Cluster56_1 NA 0.9352221617365788 3/2003 1522/2003 54112.0 Not_record
>>8241-G:69070,2107-C:58497,1512-T:55353,2779-C:51628,2458-T:48553,3256-C:42138,3694-G:39726,Genome_num: 1
>JX878305.1 Cluster94_1 NA 0.9136790223923512 2/2007 1486/2007 56925.0 Not_record
>>8241-G:69070,2779-C:51628,2458-T:48553,3256-C:42138,3694-G:39726,Genome_num: 1
>MN607197.1 Cluster101_2 NA 0.7805906271469676 2/2007 1268/2007 44139.5 Not_record
>>(Could be FP)8241-G:69070,2779-C:51628,3256-C:42138,Genome_num: 13
>OR361533.1 Cluster8_1 NA 0.9396169447153184 1/2007 1531/2007 69070.0 Not_record
>OR361522.1 Cluster17_1 Not_record 0.9371113299147774 1/2007 1527/2007 69070.0 Not_record
>OP748245.1 Cluster97_1 Not_record 0.9262635441332909 1/2007 1507/2007 69070.0 Not_record
>OR361528.1 Cluster30_1 Not_record 0.9245093723538664 1/2007 1506/2007 69070.0 Not_record
>OR361523.1 Cluster29_1 Not_record 0.9219226547492584 1/2007 1502/2007 69070.0 Not_record
>MG831200.1 Cluster21_1 Not_record 0.9206934702609288 1/2007 1497/2007 69070.0 Not_record
>MG831202.1 Cluster22_1 Not_record 0.9169451188497799 1/2007 1493/2007 69070.0 Not_record
>OR361526.1 Cluster27_1 Not_record 0.9167016005079737 1/2007 1492/2007 69070.0 Not_record
>MG831203.1 Cluster23_1 Not_record 0.9144893506676717 1/2007 1487/2007 69070.0 Not_record
>PP437982.1 Cluster209_1 Not_record 0.7110990761837801 1/2007 1154/2007 69070.0 Not_record
>PP437983.1 Cluster198_1 Not_record 0.6771802250790584 1/2007 1098/2007 69070.0 Not_record
>PP437883.1 Cluster196_1 Not_record 0.6736188393066781 1/2007 1091/2007 69070.0 Not_record
>PP437885.1 Cluster197_1 Not_record 0.6735512769621339 1/2007 1091/2007 69070.0 Not_record
>>(Could be FP)2779-C:51628,3256-C:42138,Genome_num: 1
>MG831201.1 Cluster26_1 NA 0.9171374960072556 1/2007 1493/2007 51628.0 Not_record
>>(Could be FP)3256-C:42138,Genome_num: 1
>OR361521.1 Cluster31_1 NA 0.919201512220912 1/2007 1499/2007 42138.0 Not_record

When analyzing strain composition in a sample, ​​should I first exclude strains labeled as "Could be FP" under "Other possible strains​​", ​​and then​​ calculate the relative abundance of each remaining strain based on their Strain_Depth?
Is it necessary to ​​remove host sequences​​ from the raw sequencing data ​​before running VirStrain​​ for strain detection?

[liaoherui]

Hi,

"Should I first exclude strains labeled as 'Could be FP'?" -> If your analysis prioritizes precision over recall, yes, you can exclude these strains.

"Is it necessary to remove host sequences?" -> According to Table 2 in our paper, there are very few overlapping k-mers between the host genome and the three tested viruses. However, if you found significant overlaps in your datasets, removing host sequences beforehand would be beneficial.

[shiny1110]

When calculating the relative abundance of different viral strains, can I directly use the ratio of their Strain_Depth values?

[liaoherui]

Yes.

[shiny1110]

Thank you for your patient and detailed explanatio! I truly appreciate your guidance.