数据集获取

git clone https://github.com/GenerGener/HIV-ONT-Datasets.git

curl -L -o repo.zip https://github.com/GenerGener/HIV-ONT-Datasets/archive/refs/heads/master.zip
unzip repo.zip

源站 README 摘录（使用方式）

HIV-ONT-Datasets

Alejandro Gener (ARG)
Practice dataset for native viral RNA sequencing and assembly. These data provide a model (ground truth) for direct viral haplotyping for SNVs at least 6 kb apart. Boo quasispecies. This information was made available for the COVID-19 Biohackathon April 5-11 2020.

Background

The Tg26 HIV mouse has gagpol-deleted HIV integrated artificially (inserted) into the mouse germline (Dickie, et al. 1991). These insertion sites have not been characterized at depth. A recent PCR-free gDNA sequencing run (unpublished) supported 15 HIV copies across multiple insertion sites (Table 1). At the time, that study was limited by inherrent shortcomings of short read sequencing (paired-end 150, average insert sized 511). Variants in HIV copies further apart than 511 bases would not be able to be phased with this short-read data. ONT DNA error profiles approximate native RNA profiles, and improve as new basecalling models develop (Gener, 2019; Gener and Kimata, 2019).

Method

One Tg26 mouse was humanely euthenized according to IACUC-approved protocols. Genomic DNA from the bone marrow of one Tg26 male mouse was extracted with special consideration for maintaining longer gDNA. Longrange PCR was performed to create ~6 kb amplicons. DNA was sequence on one MinION (Oxford Nanopore Technologies, Oxford, UK), with SQK-LSK109 accroding to manufacturer’'s protocol. Live basecalling with MinKNOW verion 19.12.5 was done on High-accuracy mode. Read filtering was done in Galaxy, specifying reads between 6 and 6.25 kb. Read mapping was done with minimap2 in Galaxy. Visualized in IGV. HIV-1 reference used: HIV-1 vector pNL4-3, complete sequence AF324493.2. All analyses as “base 1.” No human patient data is deposited in this page.

Results

Table 1: Variants in a Tg26 mouse
There are 49 single nucleotide variants across the expected HIV portion of the transgene, supported by deep short read sequencing. Three are adjoining (and thus linked), so there were 46 variants across 15 HIV copies in this individual mouse.
Figure 1: Subset of reads that were acquired by the end of the long amplicon sequencing run
The deletion seen toward left side is expected in this model. A few reads mapping at 3’’ LTR are likely 5’’ LTR reads with sequencing errors in key places

数据加载示例（表格/文本类）

import pandas as pd, glob, os

files = (glob.glob(os.path.join(path, "**", "*.csv"), recursive=True)
       + glob.glob(os.path.join(path, "**", "*.tsv"), recursive=True)
       + glob.glob(os.path.join(path, "**", "*.xlsx"), recursive=True))
print("数据文件:", files)
df = pd.read_csv(files[0])
print(df.shape); print(df.columns.tolist()); print(df.head(3))

完整仓库：github.com/GenerGener/HIV-ONT-Datasets