US 12,454,724 B2
Methods and systems for high-depth sequencing of methylated nucleic acid
Eric Ariazi, Redwood City, CA (US); David Weinberg, San Francisco, CA (US); Greg Hogan, San Francisco, CA (US); John St. John, Brisbane, CA (US); and Michael Pearson, San Francisco, CA (US)
Assigned to Freenome Holdings, Inc., Brisbane, CA (US)
Filed by Freenome Holdings, Inc., South San Francisco, CA (US)
Filed on Feb. 1, 2023, as Appl. No. 18/163,128.
Application 18/163,128 is a continuation of application No. 17/173,770, filed on Feb. 11, 2021.
Application 17/173,770 is a continuation of application No. PCT/US2020/035380, filed on May 29, 2020.
Claims priority of provisional application 62/855,795, filed on May 31, 2019.
Prior Publication US 2023/0323446 A1, Oct. 12, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. C12Q 1/6869 (2018.01); C12Q 1/6855 (2018.01); C12Q 1/686 (2018.01); C12Q 1/6876 (2018.01)
CPC C12Q 1/6869 (2013.01) [C12Q 1/6855 (2013.01); C12Q 1/686 (2013.01); C12Q 1/6876 (2013.01); C12Q 2600/154 (2013.01)] 16 Claims
 
1. A method of detecting a cancer in a subject, comprising:
(a) ligating a nucleic acid adapter to a nucleic acid molecule of a nucleic acid sample obtained or derived from the subject, wherein the nucleic acid molecule comprises unmethylated cytosines, wherein the nucleic acid adapter is a conversion-tolerant adapter comprising guanine, thymine, adenine, and cytosine bases, and not comprising 5-methylcytosine (5 mC)-containing bases, and not comprising 5-hydroxymethylcytosine (5 hmC)-containing bases;
(b) converting the unmethylated cytosines of the nucleic acid molecule to uracils using an enzymatic conversion method, thereby generating a converted nucleic acid molecule;
(c) amplifying the converted nucleic acid molecule at least in part by polymerase chain reaction (PCR), thereby generating amplified converted nucleic acid molecules, wherein the PCR comprises use of sequencing primers that correspond to a sequence of the conversion-tolerant adapter, wherein primers of the PCR comprise a sequence selected from the group consisting of SEQ ID NOs. 3-7 and 13;
(d) contacting the amplified converted nucleic acid molecules with nucleic acid probes that are at least partially complementary to a pre-identified panel of CpG, CHG, or CHH loci to enrich for sequences corresponding to the pre-identified panel of CpG, CHG, or CHH loci, thereby generating enriched nucleic acid molecules;
(e) sequencing the enriched nucleic acid molecules or a derivative thereof at a depth of greater than 100×; and
(f) using a computer specifically programmed to detect the cancer to perform at least:
(i) comparing the nucleic acid sequence of the enriched nucleic acid molecules or a derivative thereof to a reference nucleic acid sequence of the pre-identified panel of CpG, CHG, or CHH loci; and
(ii) processing the nucleic acid sequence of the enriched nucleic acid molecules or a derivative thereof using a trained machine learning model configured to detect the cancer in the subject based at least in part on the comparing in (i), wherein the trained machine learning model is trained with training data comprising: (1) a first set of biological samples obtained or derived from subjects with advanced adenoma, and (2) a second set of biological samples obtained or derived from subjects without advanced adenoma.