<p>With the development of next-generation sequencing (NGS) technologies it became possible to simultaneously analyze millions of variants. Despite the quality improvement, it is generally still required to confirm the variants before reporting. However, in recent years the dominant idea is that one could define the quality thresholds for “high quality” variants which do not require orthogonal validation. Despite that, no works to date report the concordance between variants from whole genome sequencing and their gold-standard Sanger validation. In this study we analyzed the concordance for 1756 WGS variants in order to establish the appropriate thresholds for high-quality variants filtering. Resulting thresholds allowed us to drastically reduce the number of variants which require validation, to 4.8% and 1.2% of the initial set for caller-agnostic (DP, AF) and caller-dependent (QUAL) thresholds, respectively.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Sanger validation of WGS variants

  • Arina Kopernik,
  • Mariia Sayganova,
  • Gaukhar Zobkova,
  • Natalia Doroschuk,
  • Anna Smirnova,
  • Daria Molodtsova-Zolotukhina,
  • Olesya Sagaydak,
  • Oxana Ryzhkova,
  • Sergey Kutsev,
  • Olga Groznova,
  • Lyusya Melikyan,
  • Elizaveta Bondarchuk,
  • Mary Woroncow,
  • Eugene Albert,
  • Viktor Bogdanov,
  • Pavel Volchkov

摘要

With the development of next-generation sequencing (NGS) technologies it became possible to simultaneously analyze millions of variants. Despite the quality improvement, it is generally still required to confirm the variants before reporting. However, in recent years the dominant idea is that one could define the quality thresholds for “high quality” variants which do not require orthogonal validation. Despite that, no works to date report the concordance between variants from whole genome sequencing and their gold-standard Sanger validation. In this study we analyzed the concordance for 1756 WGS variants in order to establish the appropriate thresholds for high-quality variants filtering. Resulting thresholds allowed us to drastically reduce the number of variants which require validation, to 4.8% and 1.2% of the initial set for caller-agnostic (DP, AF) and caller-dependent (QUAL) thresholds, respectively.